International Powerlifting Federation (IPF)¶

2018-2024 Exploratory Data Analysis¶

image.png

Credits: https://www.powerlifting.sport/¶

Powerlifting is a competitive strength sport that focuses on three primary lifts: the squat, bench press, and deadlift. Participants in powerlifting competitions, known as powerlifters, aim to lift as much weight as possible in each of these three lifts within specific weight categories and age divisions¶

About the Data¶

The powerlifting data was downloded from https://openpowerlifting.gitlab.io/opl-csv/bulk-csv.html, containing 1.2million competition records from 2002-2024.¶

Data Wrangling¶

I will start by setting up the environment, loading the dataset, and performing an initial assessment of its contents. Following that, we will proceed with cleaning and processing the data to make it suitable for analysis.¶
In [1]:
# Importing libraries
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns

pd.set_option('display.max_rows', 500)
pd.set_option('display.max_columns', 500)
pd.set_option('display.width', 1000)
In [2]:
# Read in file
df = pd.read_csv("../openipf-2024-02-24/openipf-2024-02-24.csv")

C:\Users\User\AppData\Local\Temp\ipykernel_20184\3173043467.py:2: DtypeWarning: Columns (33,38) have mixed types. Specify dtype option on import or set low_memory=False.
  df = pd.read_csv("../openipf-2024-02-24/openipf-2024-02-24.csv")
In [3]:
# See the data that's been imported
df.head()
Out[3]:
Name Sex Event Equipment Age AgeClass BirthYearClass Division BodyweightKg WeightClassKg Squat1Kg Squat2Kg Squat3Kg Squat4Kg Best3SquatKg Bench1Kg Bench2Kg Bench3Kg Bench4Kg Best3BenchKg Deadlift1Kg Deadlift2Kg Deadlift3Kg Deadlift4Kg Best3DeadliftKg TotalKg Place Dots Wilks Glossbrenner Goodlift Tested Country State Federation ParentFederation Date MeetCountry MeetState MeetTown MeetName
0 Karen Thomson F SBD Single-ply 39.5 35-39 40-49 Masters 1 51.90 52 105.0 110.5 -115.0 NaN 110.5 67.5 70.5 72.5 NaN 72.5 125.0 130.0 -136.0 NaN 130.0 312.5 1 381.42 390.15 345.86 63.73 Yes New Zealand NaN OceaniaPF IPF 2002-12-06 New Zealand BOP Tauranga Oceania Championships
1 Deborah Groves F SBD Single-ply 30.5 24-34 24-39 Open 50.64 52 122.5 -127.5 127.5 NaN 127.5 82.5 87.5 90.0 NaN 90.0 140.0 152.5 -157.0 NaN 152.5 370.0 1 459.45 470.74 417.74 76.76 Yes Australia NaN OceaniaPF IPF 2002-12-06 New Zealand BOP Tauranga Oceania Championships
2 Renee Orbon F SBD Single-ply 31.5 24-34 24-39 Open 55.76 56 105.5 110.0 115.0 NaN 115.0 82.5 85.0 87.5 NaN 87.5 127.5 132.5 137.5 NaN 137.5 340.0 1 395.22 401.40 355.08 66.09 Yes New Zealand NaN OceaniaPF IPF 2002-12-06 New Zealand BOP Tauranga Oceania Championships
3 Rikki Flanders F SBD Single-ply 31.5 24-34 24-39 Open 56.42 60 110.0 117.5 -120.0 NaN 117.5 62.5 -65.0 -65.0 NaN 62.5 130.0 137.5 140.0 NaN 140.0 320.0 1 369.09 374.31 331.05 61.74 Yes Australia NaN OceaniaPF IPF 2002-12-06 New Zealand BOP Tauranga Oceania Championships
4 Julie Hallmond F SBD Single-ply NaN NaN 50-59 Masters 2 58.20 60 105.0 110.0 112.5 NaN 112.5 57.5 60.0 62.5 NaN 62.5 135.0 137.5 141.0 NaN 141.0 316.0 1 357.17 360.76 318.88 59.79 Yes New Zealand NaN OceaniaPF IPF 2002-12-06 New Zealand BOP Tauranga Oceania Championships
In [4]:
df.shape
Out[4]:
(1228685, 41)
In [5]:
df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 1228685 entries, 0 to 1228684
Data columns (total 41 columns):
 #   Column            Non-Null Count    Dtype  
---  ------            --------------    -----  
 0   Name              1228685 non-null  object 
 1   Sex               1228685 non-null  object 
 2   Event             1228685 non-null  object 
 3   Equipment         1228685 non-null  object 
 4   Age               890357 non-null   float64
 5   AgeClass          933647 non-null   object 
 6   BirthYearClass    949116 non-null   object 
 7   Division          1227735 non-null  object 
 8   BodyweightKg      1214221 non-null  float64
 9   WeightClassKg     1215634 non-null  object 
 10  Squat1Kg          488458 non-null   float64
 11  Squat2Kg          483645 non-null   float64
 12  Squat3Kg          471823 non-null   float64
 13  Squat4Kg          327 non-null      float64
 14  Best3SquatKg      884070 non-null   float64
 15  Bench1Kg          690577 non-null   float64
 16  Bench2Kg          684269 non-null   float64
 17  Bench3Kg          666876 non-null   float64
 18  Bench4Kg          1255 non-null     float64
 19  Best3BenchKg      1151708 non-null  float64
 20  Deadlift1Kg       498206 non-null   float64
 21  Deadlift2Kg       490565 non-null   float64
 22  Deadlift3Kg       473776 non-null   float64
 23  Deadlift4Kg       474 non-null      float64
 24  Best3DeadliftKg   896962 non-null   float64
 25  TotalKg           1160339 non-null  float64
 26  Place             1228685 non-null  object 
 27  Dots              1147470 non-null  float64
 28  Wilks             1147470 non-null  float64
 29  Glossbrenner      1147470 non-null  float64
 30  Goodlift          1117856 non-null  float64
 31  Tested            1088779 non-null  object 
 32  Country           715827 non-null   object 
 33  State             221128 non-null   object 
 34  Federation        1228685 non-null  object 
 35  ParentFederation  1228685 non-null  object 
 36  Date              1228685 non-null  object 
 37  MeetCountry       1228685 non-null  object 
 38  MeetState         695460 non-null   object 
 39  MeetTown          1025789 non-null  object 
 40  MeetName          1228685 non-null  object 
dtypes: float64(22), object(19)
memory usage: 384.3+ MB

Data Preparation¶

1. Filter data only from 2018-2024¶

2. Only include events that include all 3 movements: Squat, Bench and Deadlift, aka SBD events.¶

3. Fliter competition meets that does not use any supportive equipment¶

4. Filter divisions to include: Sub Juniors, Juniors, Opens, and Masters 1-4¶

5. Clean meet results for squat, bench and deadlift¶

6. Clean weight classes¶

7. Clean other columns such as Federation, MeetState, MeetTown etc.¶

8. Removing duplicates¶

Step 1: Show 2018-2024¶

In [6]:
df['Date']
Out[6]:
0          2002-12-06
1          2002-12-06
2          2002-12-06
3          2002-12-06
4          2002-12-06
              ...    
1228680    2016-11-14
1228681    2016-11-14
1228682    2016-11-14
1228683    2016-11-14
1228684    2016-11-14
Name: Date, Length: 1228685, dtype: object
In [7]:
# Since dtype is object, convert 'Date' column to datetime type
df['Date'] = pd.to_datetime(df['Date'])
In [8]:
# Extract year from 'Date' column
df['Year'] = df['Date'].dt.year
In [9]:
# Filtering to 2020-2024
df2 = df[df['Year'].between(2018, 2024)]
In [10]:
# Sort out values
df2 = df2.sort_values(by='Year', ascending=True)
df2 = df2.reset_index(drop=True)
In [11]:
df2.head(10)
Out[11]:
Name Sex Event Equipment Age AgeClass BirthYearClass Division BodyweightKg WeightClassKg Squat1Kg Squat2Kg Squat3Kg Squat4Kg Best3SquatKg Bench1Kg Bench2Kg Bench3Kg Bench4Kg Best3BenchKg Deadlift1Kg Deadlift2Kg Deadlift3Kg Deadlift4Kg Best3DeadliftKg TotalKg Place Dots Wilks Glossbrenner Goodlift Tested Country State Federation ParentFederation Date MeetCountry MeetState MeetTown MeetName Year
0 Taran Dhanju F SBD Raw 44.5 40-44 40-49 Masters 1 63.0 63 NaN NaN NaN NaN 90.0 NaN NaN NaN NaN 40.0 NaN NaN NaN NaN 142.5 272.5 1 293.08 292.65 258.42 59.62 Yes Canada ON CPU IPF 2018-06-09 Canada ON Niagara Niagara Open 2018
1 Patrick Mitchell M BD Single-ply 41.5 40-44 40-49 M-O 115.2 120 NaN NaN NaN NaN NaN -200.0 -200.0 -200.0 NaN NaN 205.0 215.0 -227.5 NaN 215.0 NaN DQ NaN NaN NaN NaN Yes USA MI USAPL IPF 2018-10-13 USA MI NaN Michigan Iron Monger 2018
2 Jeffrey Skorupa M BD Single-ply 52.5 50-54 50-59 M-M2a 121.4 120+ NaN NaN NaN NaN NaN 220.0 232.5 242.5 NaN 242.5 215.0 230.0 245.0 NaN 245.0 487.5 1 278.90 279.53 267.86 NaN Yes USA MI USAPL IPF 2018-10-13 USA MI NaN Michigan Iron Monger 2018
3 Janan Khalil F BD Raw 15.5 16-17 14-18 FR-T2 84.7 84+ NaN NaN NaN NaN NaN 55.0 62.5 -65.0 NaN 62.5 137.5 152.5 160.0 NaN 160.0 222.5 1 203.93 197.60 172.65 NaN Yes USA MI USAPL IPF 2018-10-13 USA MI NaN Michigan Iron Monger 2018
4 Megan Sisson F BD Raw 15.5 16-17 14-18 FR-T2 68.5 72 NaN NaN NaN NaN NaN 55.0 -57.5 -57.5 NaN 55.0 122.5 127.5 132.5 NaN 132.5 187.5 1 191.95 189.37 167.01 NaN Yes USA MI USAPL IPF 2018-10-13 USA MI NaN Michigan Iron Monger 2018
5 Angelina Selou F BD Raw 19.5 20-23 19-23 FR-O 88.4 84+ NaN NaN NaN NaN NaN 40.0 -45.0 -45.0 NaN 40.0 85.0 92.5 100.0 NaN 100.0 140.0 5 125.82 121.90 106.13 NaN Yes USA MI USAPL IPF 2018-10-13 USA MI NaN Michigan Iron Monger 2018
6 Lana Goodrich F BD Raw 34.5 24-34 24-39 FR-O 107.1 84+ NaN NaN NaN NaN NaN 55.0 -60.0 60.0 NaN 60.0 107.5 NaN NaN NaN 107.5 167.5 4 139.35 136.99 116.72 NaN Yes USA MI USAPL IPF 2018-10-13 USA MI NaN Michigan Iron Monger 2018
7 Noel Gray F BD Raw 45.5 45-49 40-49 FR-O 142.7 84+ NaN NaN NaN NaN NaN 57.5 62.5 -70.0 NaN 62.5 105.0 112.5 120.0 NaN 120.0 182.5 3 141.24 141.44 119.03 NaN Yes USA MI USAPL IPF 2018-10-13 USA MI NaN Michigan Iron Monger 2018
8 Janan Khalil F BD Raw 15.5 16-17 14-18 FR-O 84.7 84+ NaN NaN NaN NaN NaN 55.0 62.5 -65.0 NaN 62.5 137.5 152.5 160.0 NaN 160.0 222.5 2 203.93 197.60 172.65 NaN Yes USA MI USAPL IPF 2018-10-13 USA MI NaN Michigan Iron Monger 2018
9 Susan Mattingly F BD Raw 52.0 50-54 50-59 FR-O 99.6 84+ NaN NaN NaN NaN NaN 80.0 95.0 -100.0 NaN 95.0 127.5 145.0 152.5 NaN 152.5 247.5 1 211.53 206.31 177.32 NaN Yes USA MI USAPL IPF 2018-10-13 USA MI NaN Michigan Iron Monger 2018
In [12]:
df2.shape
Out[12]:
(390850, 42)

Step 2: Filter events that include all 3 movements: Squat, Bench and Deadlift (SBD events)¶

In [13]:
df2['Event'].unique()
Out[13]:
array(['SBD', 'BD', 'D', 'B', 'SB', 'S', 'SD'], dtype=object)
In [14]:
df2[df2['Event'].str.contains('SBD')]
Out[14]:
Name Sex Event Equipment Age AgeClass BirthYearClass Division BodyweightKg WeightClassKg Squat1Kg Squat2Kg Squat3Kg Squat4Kg Best3SquatKg Bench1Kg Bench2Kg Bench3Kg Bench4Kg Best3BenchKg Deadlift1Kg Deadlift2Kg Deadlift3Kg Deadlift4Kg Best3DeadliftKg TotalKg Place Dots Wilks Glossbrenner Goodlift Tested Country State Federation ParentFederation Date MeetCountry MeetState MeetTown MeetName Year
0 Taran Dhanju F SBD Raw 44.5 40-44 40-49 Masters 1 63.00 63 NaN NaN NaN NaN 90.0 NaN NaN NaN NaN 40.0 NaN NaN NaN NaN 142.5 272.5 1 293.08 292.65 258.42 59.62 Yes Canada ON CPU IPF 2018-06-09 Canada ON Niagara Niagara Open 2018
53 Bryan Sledge M SBD Raw 18.5 18-19 19-23 MR-T3 81.10 83 110.0 125.0 132.5 NaN 132.5 -100.0 100.0 NaN NaN 100.0 175.0 195.0 205.0 NaN 205.0 437.5 1 299.28 296.15 285.17 61.28 Yes USA AK USAPL IPF 2018-02-24 USA NE NaN Capital City Clash 2018
54 Robert Stolley M SBD Raw 26.5 24-34 24-39 MR-O 212.00 120+ -172.5 172.5 190.0 NaN 190.0 -105.0 -105.0 -105.0 NaN NaN -215.0 -215.0 -215.0 NaN NaN NaN DQ NaN NaN NaN NaN Yes USA NE USAPL IPF 2018-02-24 USA NE NaN Capital City Clash 2018
55 Patrick Mcconnell M SBD Raw 22.5 20-23 19-23 MR-O 125.50 120+ 150.0 165.0 -177.5 NaN 165.0 115.0 125.0 127.5 NaN 127.5 225.0 242.5 252.5 NaN 252.5 545.0 2 308.49 310.32 296.85 62.14 Yes USA NE USAPL IPF 2018-02-24 USA NE NaN Capital City Clash 2018
56 Mike Kracl M SBD Raw 29.5 24-34 24-39 MR-O 126.80 120+ 247.5 262.5 272.5 NaN 272.5 162.5 170.0 -177.5 NaN 170.0 257.5 272.5 285.0 NaN 285.0 727.5 1 410.47 413.39 395.42 82.59 Yes USA NE USAPL IPF 2018-02-24 USA NE NaN Capital City Clash 2018
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
390845 Elise Turcotte F SBD Raw NaN NaN NaN Open 91.89 84+ 82.5 87.5 95.0 NaN 95.0 47.5 -52.5 55.0 NaN 55.0 120.0 130.0 135.0 NaN 135.0 285.0 5 251.79 244.21 211.84 52.12 Yes Canada AB CPU IPF 2024-01-20 Canada AB Edmonton Grit Power Open 2024
390846 David Dacong M SBD Raw NaN 20-23 19-23 Juniors 65.13 66 160.0 172.5 177.5 NaN 177.5 105.0 -115.0 -120.0 NaN 105.0 170.0 175.0 182.5 NaN 182.5 465.0 1 367.97 369.15 358.94 73.00 Yes Canada AB CPU IPF 2024-01-20 Canada AB Edmonton Grit Power Open 2024
390847 Alexander Dubois M SBD Raw NaN NaN 19-23 Juniors 64.05 66 150.0 157.5 -162.5 NaN 157.5 -95.0 95.0 97.5 NaN 97.5 -180.0 185.0 195.0 NaN 195.0 450.0 2 360.64 362.31 352.58 71.27 Yes Canada AB CPU IPF 2024-01-20 Canada AB Edmonton Grit Power Open 2024
390848 Edward Boraas M SBD Raw NaN NaN 40-49 Masters 1 81.28 83 157.5 162.5 170.0 NaN 170.0 110.0 112.5 -115.0 NaN 112.5 185.0 195.0 205.0 NaN 205.0 487.5 1 333.05 329.54 317.30 68.20 Yes Canada AB CPU IPF 2024-01-20 Canada AB Edmonton Grit Power Open 2024
390849 Maria Saldana F SBD Raw NaN NaN NaN Sub-Juniors 71.36 76 107.5 112.5 -120.0 NaN 112.5 57.5 62.5 -65.0 NaN 62.5 115.0 120.0 127.5 NaN 127.5 302.5 1 302.65 297.02 261.69 61.54 Yes Canada NS CPU IPF 2024-02-10 Canada NS Halifax Frozen Fury 2024

302024 rows × 42 columns

In [15]:
# Update Dataframe
df2 = df2[df2['Event'].str.contains('SBD')]

Step 3: Filter competition meets that does not use any supportive equipment¶

In [16]:
df2['Equipment'].unique()
Out[16]:
array(['Raw', 'Single-ply', 'Wraps'], dtype=object)
In [17]:
df2[df2['Equipment'].str.contains('Raw')]
Out[17]:
Name Sex Event Equipment Age AgeClass BirthYearClass Division BodyweightKg WeightClassKg Squat1Kg Squat2Kg Squat3Kg Squat4Kg Best3SquatKg Bench1Kg Bench2Kg Bench3Kg Bench4Kg Best3BenchKg Deadlift1Kg Deadlift2Kg Deadlift3Kg Deadlift4Kg Best3DeadliftKg TotalKg Place Dots Wilks Glossbrenner Goodlift Tested Country State Federation ParentFederation Date MeetCountry MeetState MeetTown MeetName Year
0 Taran Dhanju F SBD Raw 44.5 40-44 40-49 Masters 1 63.00 63 NaN NaN NaN NaN 90.0 NaN NaN NaN NaN 40.0 NaN NaN NaN NaN 142.5 272.5 1 293.08 292.65 258.42 59.62 Yes Canada ON CPU IPF 2018-06-09 Canada ON Niagara Niagara Open 2018
53 Bryan Sledge M SBD Raw 18.5 18-19 19-23 MR-T3 81.10 83 110.0 125.0 132.5 NaN 132.5 -100.0 100.0 NaN NaN 100.0 175.0 195.0 205.0 NaN 205.0 437.5 1 299.28 296.15 285.17 61.28 Yes USA AK USAPL IPF 2018-02-24 USA NE NaN Capital City Clash 2018
54 Robert Stolley M SBD Raw 26.5 24-34 24-39 MR-O 212.00 120+ -172.5 172.5 190.0 NaN 190.0 -105.0 -105.0 -105.0 NaN NaN -215.0 -215.0 -215.0 NaN NaN NaN DQ NaN NaN NaN NaN Yes USA NE USAPL IPF 2018-02-24 USA NE NaN Capital City Clash 2018
55 Patrick Mcconnell M SBD Raw 22.5 20-23 19-23 MR-O 125.50 120+ 150.0 165.0 -177.5 NaN 165.0 115.0 125.0 127.5 NaN 127.5 225.0 242.5 252.5 NaN 252.5 545.0 2 308.49 310.32 296.85 62.14 Yes USA NE USAPL IPF 2018-02-24 USA NE NaN Capital City Clash 2018
56 Mike Kracl M SBD Raw 29.5 24-34 24-39 MR-O 126.80 120+ 247.5 262.5 272.5 NaN 272.5 162.5 170.0 -177.5 NaN 170.0 257.5 272.5 285.0 NaN 285.0 727.5 1 410.47 413.39 395.42 82.59 Yes USA NE USAPL IPF 2018-02-24 USA NE NaN Capital City Clash 2018
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
390845 Elise Turcotte F SBD Raw NaN NaN NaN Open 91.89 84+ 82.5 87.5 95.0 NaN 95.0 47.5 -52.5 55.0 NaN 55.0 120.0 130.0 135.0 NaN 135.0 285.0 5 251.79 244.21 211.84 52.12 Yes Canada AB CPU IPF 2024-01-20 Canada AB Edmonton Grit Power Open 2024
390846 David Dacong M SBD Raw NaN 20-23 19-23 Juniors 65.13 66 160.0 172.5 177.5 NaN 177.5 105.0 -115.0 -120.0 NaN 105.0 170.0 175.0 182.5 NaN 182.5 465.0 1 367.97 369.15 358.94 73.00 Yes Canada AB CPU IPF 2024-01-20 Canada AB Edmonton Grit Power Open 2024
390847 Alexander Dubois M SBD Raw NaN NaN 19-23 Juniors 64.05 66 150.0 157.5 -162.5 NaN 157.5 -95.0 95.0 97.5 NaN 97.5 -180.0 185.0 195.0 NaN 195.0 450.0 2 360.64 362.31 352.58 71.27 Yes Canada AB CPU IPF 2024-01-20 Canada AB Edmonton Grit Power Open 2024
390848 Edward Boraas M SBD Raw NaN NaN 40-49 Masters 1 81.28 83 157.5 162.5 170.0 NaN 170.0 110.0 112.5 -115.0 NaN 112.5 185.0 195.0 205.0 NaN 205.0 487.5 1 333.05 329.54 317.30 68.20 Yes Canada AB CPU IPF 2024-01-20 Canada AB Edmonton Grit Power Open 2024
390849 Maria Saldana F SBD Raw NaN NaN NaN Sub-Juniors 71.36 76 107.5 112.5 -120.0 NaN 112.5 57.5 62.5 -65.0 NaN 62.5 115.0 120.0 127.5 NaN 127.5 302.5 1 302.65 297.02 261.69 61.54 Yes Canada NS CPU IPF 2024-02-10 Canada NS Halifax Frozen Fury 2024

267460 rows × 42 columns

In [18]:
# Update Dataframe
df2 = df2[df2['Equipment'].str.contains('Raw')]
In [19]:
df2.head()
Out[19]:
Name Sex Event Equipment Age AgeClass BirthYearClass Division BodyweightKg WeightClassKg Squat1Kg Squat2Kg Squat3Kg Squat4Kg Best3SquatKg Bench1Kg Bench2Kg Bench3Kg Bench4Kg Best3BenchKg Deadlift1Kg Deadlift2Kg Deadlift3Kg Deadlift4Kg Best3DeadliftKg TotalKg Place Dots Wilks Glossbrenner Goodlift Tested Country State Federation ParentFederation Date MeetCountry MeetState MeetTown MeetName Year
0 Taran Dhanju F SBD Raw 44.5 40-44 40-49 Masters 1 63.0 63 NaN NaN NaN NaN 90.0 NaN NaN NaN NaN 40.0 NaN NaN NaN NaN 142.5 272.5 1 293.08 292.65 258.42 59.62 Yes Canada ON CPU IPF 2018-06-09 Canada ON Niagara Niagara Open 2018
53 Bryan Sledge M SBD Raw 18.5 18-19 19-23 MR-T3 81.1 83 110.0 125.0 132.5 NaN 132.5 -100.0 100.0 NaN NaN 100.0 175.0 195.0 205.0 NaN 205.0 437.5 1 299.28 296.15 285.17 61.28 Yes USA AK USAPL IPF 2018-02-24 USA NE NaN Capital City Clash 2018
54 Robert Stolley M SBD Raw 26.5 24-34 24-39 MR-O 212.0 120+ -172.5 172.5 190.0 NaN 190.0 -105.0 -105.0 -105.0 NaN NaN -215.0 -215.0 -215.0 NaN NaN NaN DQ NaN NaN NaN NaN Yes USA NE USAPL IPF 2018-02-24 USA NE NaN Capital City Clash 2018
55 Patrick Mcconnell M SBD Raw 22.5 20-23 19-23 MR-O 125.5 120+ 150.0 165.0 -177.5 NaN 165.0 115.0 125.0 127.5 NaN 127.5 225.0 242.5 252.5 NaN 252.5 545.0 2 308.49 310.32 296.85 62.14 Yes USA NE USAPL IPF 2018-02-24 USA NE NaN Capital City Clash 2018
56 Mike Kracl M SBD Raw 29.5 24-34 24-39 MR-O 126.8 120+ 247.5 262.5 272.5 NaN 272.5 162.5 170.0 -177.5 NaN 170.0 257.5 272.5 285.0 NaN 285.0 727.5 1 410.47 413.39 395.42 82.59 Yes USA NE USAPL IPF 2018-02-24 USA NE NaN Capital City Clash 2018

Step 4: Filter age divisions to include: Sub Juniors, Juniors, Opens, and Masters 1-3¶

This is what i gathered from the ipf technical rules book (pg.3,4)¶

Credits: https://www.powerlifting.sport/fileadmin/ipf/data/rules/technical-rules/english/IPF_Technical_Rules_Book_2021docx.pdf¶

image.png

In [20]:
df2['BirthYearClass'].unique()
Out[20]:
array(['40-49', '19-23', '24-39', '14-18', '50-59', nan, '70-999',
       '60-69'], dtype=object)
In [21]:
# Filtering null values for birthyearclass 
df2[df2['BirthYearClass'].isnull()]
Out[21]:
Name Sex Event Equipment Age AgeClass BirthYearClass Division BodyweightKg WeightClassKg Squat1Kg Squat2Kg Squat3Kg Squat4Kg Best3SquatKg Bench1Kg Bench2Kg Bench3Kg Bench4Kg Best3BenchKg Deadlift1Kg Deadlift2Kg Deadlift3Kg Deadlift4Kg Best3DeadliftKg TotalKg Place Dots Wilks Glossbrenner Goodlift Tested Country State Federation ParentFederation Date MeetCountry MeetState MeetTown MeetName Year
129 Logan Auerbach M SBD Raw 10.5 5-12 NaN MR-Y2 47.90 48 47.5 52.5 57.5 NaN 57.5 37.5 40.0 42.5 NaN 42.5 75.0 80.0 82.5 NaN 82.5 182.5 1 189.55 195.79 193.67 33.78 Yes USA OH USAPL IPF 2018-08-04 USA OH NaN Dog Days of Summer Powerlifting 2018
200 Svetlana Burko F SBD Raw 12.5 13-15 NaN Sub-Juniors 31.05 43 NaN NaN NaN NaN 47.5 NaN NaN NaN NaN 32.5 NaN NaN NaN NaN 65.0 145.0 5 215.30 239.31 233.52 NaN Yes NaN NaN FPR IPF 2018-04-18 Russia TUL Tula Russian Sports Schools Classic Powerlifting Ch... 2018
202 Arina Netrusova F SBD Raw 12.5 13-15 NaN Sub-Juniors 41.75 43 NaN NaN NaN NaN 65.0 NaN NaN NaN NaN 35.0 NaN NaN NaN NaN 75.0 175.0 3 251.10 254.86 230.12 55.13 Yes NaN NaN FPR IPF 2018-04-18 Russia TUL Tula Russian Sports Schools Classic Powerlifting Ch... 2018
203 Darya Antonova F SBD Raw 12.5 13-15 NaN Sub-Juniors 41.95 43 NaN NaN NaN NaN 85.0 NaN NaN NaN NaN 55.0 NaN NaN NaN NaN 105.0 245.0 1 350.21 355.75 321.01 76.75 Yes Russia NaN FPR IPF 2018-04-18 Russia TUL Tula Russian Sports Schools Classic Powerlifting Ch... 2018
268 Jessica DeSousa F SBD Raw NaN NaN NaN Open 60.20 63 NaN NaN NaN NaN 110.0 NaN NaN NaN NaN 60.0 NaN NaN NaN NaN 140.0 310.0 3 342.93 344.73 304.57 69.91 Yes NaN NaN VGPF IPF 2018-09-22 Belgium NaN Merchtem Vlaams Kampioenschap - Powerlifting & Bench Pr... 2018
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
390842 Nicole Pepper F SBD Raw NaN NaN NaN Open 100.51 84+ -107.5 110.0 120.0 NaN 120.0 70.0 77.5 82.5 NaN 82.5 130.0 140.0 145.0 NaN 145.0 347.5 2 295.95 288.90 248.02 61.89 Yes Canada AB CPU IPF 2024-01-20 Canada AB Edmonton Grit Power Open 2024
390843 Rebecca Krampl F SBD Raw NaN NaN NaN Open 89.80 84+ 92.5 105.0 117.5 NaN 117.5 -40.0 47.5 -55.0 NaN 47.5 115.0 -130.0 140.0 NaN 140.0 305.0 3 272.18 263.78 229.33 56.21 Yes Canada AB CPU IPF 2024-01-20 Canada AB Edmonton Grit Power Open 2024
390844 Shelby Mildenberger F SBD Raw NaN NaN NaN Open 112.21 84+ 110.0 120.0 -127.5 NaN 120.0 57.5 62.5 -65.0 NaN 62.5 110.0 120.0 -130.0 NaN 120.0 302.5 4 247.72 244.95 208.32 52.51 Yes Canada AB CPU IPF 2024-01-20 Canada AB Edmonton Grit Power Open 2024
390845 Elise Turcotte F SBD Raw NaN NaN NaN Open 91.89 84+ 82.5 87.5 95.0 NaN 95.0 47.5 -52.5 55.0 NaN 55.0 120.0 130.0 135.0 NaN 135.0 285.0 5 251.79 244.21 211.84 52.12 Yes Canada AB CPU IPF 2024-01-20 Canada AB Edmonton Grit Power Open 2024
390849 Maria Saldana F SBD Raw NaN NaN NaN Sub-Juniors 71.36 76 107.5 112.5 -120.0 NaN 112.5 57.5 62.5 -65.0 NaN 62.5 115.0 120.0 127.5 NaN 127.5 302.5 1 302.65 297.02 261.69 61.54 Yes Canada NS CPU IPF 2024-02-10 Canada NS Halifax Frozen Fury 2024

34288 rows × 42 columns

In [22]:
#34288 null values is a significant amount, let's try filtering using divisions
df2['Division'].unique()
Out[22]:
array(['Masters 1', 'MR-T3', 'MR-O', 'MR-M1b', 'MR-M1a', 'MR-T2', 'MR-Jr',
       'FR-T3', 'FR-T1', 'FR-O', 'MR-Y2', 'FR-Jr', 'Sub-Juniors',
       'FR-M2a', 'Juniors', 'Open', 'Masters 4', 'FR-T2', 'Masters 2',
       'Masters 3', 'FR-M1a', 'MR-M3a', 'MR-M4a', 'MR-M3b', 'FR-Y3',
       'FR-M1b', 'FR-M', 'FR-G', 'FR-T', 'MR-T', 'MR-M', 'MR-G', 'MR-M2b',
       'FR-M3a', 'MR-T1', 'High School', 'MR-M1', 'MR-C', 'MR-M2',
       'MR-PF', 'MR-M4', 'MR-M3', 'FR-M3', 'FR-M2', 'FR-M1', 'FR-M3b',
       'FR-Y2', 'FR-M2b', 'MR-M2a', 'FR-Y1', 'MR-M5a', 'FR-C', 'FR-M6a',
       'Juniorki do lat 18', 'Juniorki do lat 23', 'Juniorzy do lat 23',
       'MR-Y3', 'MR-Y1', 'Juniorzy do lat 18', 'FR-Y', 'MR-MO', 'MR-SO',
       'MR-Y', 'T18-19', 'M50-54', 'M45-49', 'T16-17', 'MR-HS', 'FR-HS',
       'FR-M4', 'Youth 10-13', 'Teen 14-15', 'Teen 16-17', 'M40-44',
       'M55-59', 'M-C-Open', 'M-C-U23', 'M-C-U18', 'F-C-M1', 'F-C-Open',
       'F-C-U23', 'F-C-U18', 'M-C-M1', 'M-C-M2', 'MR-JV', 'FR-JV', 'MR-V',
       'FR-M5', 'MR-TJ', 'FR-TJ', 'FR-M4a', 'Student', 'FR-M5a', 'M-C-G',
       'M-C-M4', 'M-C-M3', 'M-C-M5', 'F-C-M2', 'F-C-G', 'F-C-M3', 'Youth',
       'Youth A', 'Youth B', 'M1', 'M2', 'FR-PF', 'Seniorzy',
       'Weterani 40-49 lat', 'Weterani 50-59 lat', 'Weterani 60-69 lat',
       'MR-M4b', 'Juniorzy do lat 20', 'Seniorki', 'Juniorki do lat 20',
       'Weteranki', 'FR-V', 'Special Olympics', 'Teen 17-18',
       'Teen Under 17', 'FR-SO', 'FR-SJ', 'M3', 'M4', 'J', 'SJ', 'O',
       'M-C-U20', 'F-C-U20', 'Teen', 'MR-SJ', 'Guest', 'FR-MO', 'Junior',
       'Master I', 'Sub-Junior', 'SNR', 'MR-M5', 'Seniors', 'Cadet',
       'FR-M4b', 'Master II', 'Master III', 'JUN', 'SBJ', 'D',
       'Masters 60-69', 'Masters 50-59', 'Juniors 19-23', 'Masters 40-49',
       'Teen 14-18', 'Masters 70+', 'Under 23', 'Over 50', 'Over 40',
       'Over 60', 'Over 70', 'Under 18', 'Subjunior', 'M-O', 'F-G', 'F-O',
       'M-G', 'M-U21', 'M-U23', 'M-M1', 'M-M2', 'M-M3', 'F-U23', 'F-M2',
       'F-M3', 'F-U21', 'F-M1', 'M-E-Open', 'Under 14', 'Masters 5',
       'F-C-N', 'M-C-N', 'Novice', 'Master IV', 'SO', 'Junior I',
       'Junior II', 'Juniorki do lat 16', 'Juniorzy do lat 16',
       'Weteranki +50 lat', 'Weteranki 40-49 lat', 'Weterani +50 lat',
       'MR-J', 'Masters I', 'Masters II', 'FR-J', 'MR-Sj', 'FR-Sj',
       'Teen 15-16', 'Youth 10-14', 'FR-M6', 'Master 2', 'Master 1',
       'M-E-M2', 'Masters 1/2/3', 'Under 20', 'LW', 'HW', 'M-JR', 'F-JR',
       'Rekrutt 2', 'Prime Time', 'Especial', 'Senior', 'XR-O',
       'Weteranki 50-59 lat', 'M-SHW-U', 'M-SJ-U', 'M-J-U', 'M-M1-U',
       'M-O-U', 'F-M3-U', 'F-M2-U', 'F-M1-U', 'F-O-U', 'F-SHW-U',
       'M-LW-U', 'F-MW-U', 'M-MW-U', 'M-HW-U', 'F-LW-U', 'M-M2-U', 'Team',
       'Newcomer', 'Classic', 'Youth 2', 'Youth 3', 'F-SJ-U', 'F-J-U',
       'MR-AA', 'F-Jr', 'M-Jr', 'M-Sj', 'XR-Jr', 'M-CL-PL', 'F-CL-PL',
       'Weteranki 60-69 lat', 'University', 'Sub Junior', 'Master 3',
       'M-C-U16', 'JR', 'Junioren', 'AK I', 'AK II', 'A-Jugend',
       'Weterani +70 lat', "Men's Raw Open", "Men's Raw Junior",
       "Women's Raw Open", "Men's Raw Master III",
       "Women's Raw Sub-Junior", "Women's Raw Junior",
       "Men's Raw Sub-Junior", "Women's Raw Master III",
       "Men's Raw Guest", "Women's Raw Master I", "Women's Raw Master II",
       "Women's Raw Guest", 'M-Novice', 'F-Novice', 'M-M4', 'Youth 1',
       'Adaptive', 'MR-Pre-Sub-Jr', 'FR-Pre-Sub-Jr', 'AK', 'Under 17',
       'Developmental', 'Master', 'HS', 'C', 'G', 'MR-PS', '30+', '22-29',
       '16-22', '16-21', '22+', 'Masters +70', 'FR-PS',
       "Men's Raw Master I", 'SOI', 'Master 4', 'F-Sj', 'M-SJ', 'F-SJ'],
      dtype=object)
In [23]:
df2['AgeClass'].unique()
Out[23]:
array(['40-44', '18-19', '24-34', '20-23', '35-39', '45-49', '16-17',
       '13-15', '50-54', '5-12', '55-59', nan, '60-64', '70-74', '65-69',
       '75-79', '80-999'], dtype=object)

It would seem federations name their age divisions based on their justification rather than the IPF standards¶

Though not ideal, the best way is to remove the BirthYearClass data that contains null values as it would be impossible to justify age by Division or AgeClass. As examples, MR-T2, MR-C, HS, Prime Time are not indicative of age divisions but rather represent other categorizations such as weight classes or special divisions within the competition. Furthermore, the AgeClass contains values of varying degree, not following the IPF standards as well. Therefore, removing the rows where the 'BirthYearClass' data contains null values would be the most appropriate approach for ensuring consistency in age-based divisions across the dataset.¶

In [24]:
df2 = df2.dropna(subset=['BirthYearClass'])
In [25]:
# Check if NaN values are removed
df2['BirthYearClass'].unique()
Out[25]:
array(['40-49', '19-23', '24-39', '14-18', '50-59', '70-999', '60-69'],
      dtype=object)

Now that's completed I will have to categorize each BirthYearClass to each original ipf rule book.¶

In [26]:
# Create a dict to map each age range
age_cat_mapping = {
    '14-18': 'Sub-Junior',
    '19-23': 'Junior',
    '24-39': 'Opens',
    '40-49': 'Master 1',
    '50-59': 'Master 2',
    '60-69': 'Master 3',
    '70-999': 'Master 4',
}
# Map the age categories to the 'BirthYearClass' column
df2['AgeDivision'] = df2['BirthYearClass'].map(age_cat_mapping)
In [27]:
df2.head()
Out[27]:
Name Sex Event Equipment Age AgeClass BirthYearClass Division BodyweightKg WeightClassKg Squat1Kg Squat2Kg Squat3Kg Squat4Kg Best3SquatKg Bench1Kg Bench2Kg Bench3Kg Bench4Kg Best3BenchKg Deadlift1Kg Deadlift2Kg Deadlift3Kg Deadlift4Kg Best3DeadliftKg TotalKg Place Dots Wilks Glossbrenner Goodlift Tested Country State Federation ParentFederation Date MeetCountry MeetState MeetTown MeetName Year AgeDivision
0 Taran Dhanju F SBD Raw 44.5 40-44 40-49 Masters 1 63.0 63 NaN NaN NaN NaN 90.0 NaN NaN NaN NaN 40.0 NaN NaN NaN NaN 142.5 272.5 1 293.08 292.65 258.42 59.62 Yes Canada ON CPU IPF 2018-06-09 Canada ON Niagara Niagara Open 2018 Master 1
53 Bryan Sledge M SBD Raw 18.5 18-19 19-23 MR-T3 81.1 83 110.0 125.0 132.5 NaN 132.5 -100.0 100.0 NaN NaN 100.0 175.0 195.0 205.0 NaN 205.0 437.5 1 299.28 296.15 285.17 61.28 Yes USA AK USAPL IPF 2018-02-24 USA NE NaN Capital City Clash 2018 Junior
54 Robert Stolley M SBD Raw 26.5 24-34 24-39 MR-O 212.0 120+ -172.5 172.5 190.0 NaN 190.0 -105.0 -105.0 -105.0 NaN NaN -215.0 -215.0 -215.0 NaN NaN NaN DQ NaN NaN NaN NaN Yes USA NE USAPL IPF 2018-02-24 USA NE NaN Capital City Clash 2018 Opens
55 Patrick Mcconnell M SBD Raw 22.5 20-23 19-23 MR-O 125.5 120+ 150.0 165.0 -177.5 NaN 165.0 115.0 125.0 127.5 NaN 127.5 225.0 242.5 252.5 NaN 252.5 545.0 2 308.49 310.32 296.85 62.14 Yes USA NE USAPL IPF 2018-02-24 USA NE NaN Capital City Clash 2018 Junior
56 Mike Kracl M SBD Raw 29.5 24-34 24-39 MR-O 126.8 120+ 247.5 262.5 272.5 NaN 272.5 162.5 170.0 -177.5 NaN 170.0 257.5 272.5 285.0 NaN 285.0 727.5 1 410.47 413.39 395.42 82.59 Yes USA NE USAPL IPF 2018-02-24 USA NE NaN Capital City Clash 2018 Opens
In [28]:
# Now that Age, AgeClass, BirthYearClass, and Division are not needed, lets remove these columns
# I will also shift columns
df2 = df2[['Name', 'Sex', 'Event', 'Equipment', 
           #'Age', 
           #'AgeClass', 
           'AgeDivision', 
           #'BirthYearClass', 'Division'
           'BodyweightKg', 'WeightClassKg', 
           'Squat1Kg', 'Squat2Kg', 'Squat3Kg', 'Squat4Kg', 'Best3SquatKg', 
           'Bench1Kg', 'Bench2Kg', 'Bench3Kg', 'Bench4Kg', 'Best3BenchKg',
           'Deadlift1Kg', 'Deadlift2Kg', 'Deadlift3Kg', 'Deadlift4Kg', 'Best3DeadliftKg', 
           'TotalKg', 'Place', 'Dots', 'Wilks', 'Glossbrenner', 'Goodlift', 'Tested', 
           'Country', 'State', 'Federation', 'ParentFederation', 'Date', 'Year', 'MeetCountry', 
           'MeetState', 'MeetTown', 'MeetName']].copy()

df2 = df2.reset_index(drop=True)
In [29]:
df2.head()
Out[29]:
Name Sex Event Equipment AgeDivision BodyweightKg WeightClassKg Squat1Kg Squat2Kg Squat3Kg Squat4Kg Best3SquatKg Bench1Kg Bench2Kg Bench3Kg Bench4Kg Best3BenchKg Deadlift1Kg Deadlift2Kg Deadlift3Kg Deadlift4Kg Best3DeadliftKg TotalKg Place Dots Wilks Glossbrenner Goodlift Tested Country State Federation ParentFederation Date Year MeetCountry MeetState MeetTown MeetName
0 Taran Dhanju F SBD Raw Master 1 63.0 63 NaN NaN NaN NaN 90.0 NaN NaN NaN NaN 40.0 NaN NaN NaN NaN 142.5 272.5 1 293.08 292.65 258.42 59.62 Yes Canada ON CPU IPF 2018-06-09 2018 Canada ON Niagara Niagara Open
1 Bryan Sledge M SBD Raw Junior 81.1 83 110.0 125.0 132.5 NaN 132.5 -100.0 100.0 NaN NaN 100.0 175.0 195.0 205.0 NaN 205.0 437.5 1 299.28 296.15 285.17 61.28 Yes USA AK USAPL IPF 2018-02-24 2018 USA NE NaN Capital City Clash
2 Robert Stolley M SBD Raw Opens 212.0 120+ -172.5 172.5 190.0 NaN 190.0 -105.0 -105.0 -105.0 NaN NaN -215.0 -215.0 -215.0 NaN NaN NaN DQ NaN NaN NaN NaN Yes USA NE USAPL IPF 2018-02-24 2018 USA NE NaN Capital City Clash
3 Patrick Mcconnell M SBD Raw Junior 125.5 120+ 150.0 165.0 -177.5 NaN 165.0 115.0 125.0 127.5 NaN 127.5 225.0 242.5 252.5 NaN 252.5 545.0 2 308.49 310.32 296.85 62.14 Yes USA NE USAPL IPF 2018-02-24 2018 USA NE NaN Capital City Clash
4 Mike Kracl M SBD Raw Opens 126.8 120+ 247.5 262.5 272.5 NaN 272.5 162.5 170.0 -177.5 NaN 170.0 257.5 272.5 285.0 NaN 285.0 727.5 1 410.47 413.39 395.42 82.59 Yes USA NE USAPL IPF 2018-02-24 2018 USA NE NaN Capital City Clash
In [30]:
df2.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 233172 entries, 0 to 233171
Data columns (total 39 columns):
 #   Column            Non-Null Count   Dtype         
---  ------            --------------   -----         
 0   Name              233172 non-null  object        
 1   Sex               233172 non-null  object        
 2   Event             233172 non-null  object        
 3   Equipment         233172 non-null  object        
 4   AgeDivision       233172 non-null  object        
 5   BodyweightKg      232702 non-null  float64       
 6   WeightClassKg     231454 non-null  object        
 7   Squat1Kg          201077 non-null  float64       
 8   Squat2Kg          199716 non-null  float64       
 9   Squat3Kg          196419 non-null  float64       
 10  Squat4Kg          0 non-null       float64       
 11  Best3SquatKg      229187 non-null  float64       
 12  Bench1Kg          200652 non-null  float64       
 13  Bench2Kg          199458 non-null  float64       
 14  Bench3Kg          195488 non-null  float64       
 15  Bench4Kg          0 non-null       float64       
 16  Best3BenchKg      228386 non-null  float64       
 17  Deadlift1Kg       200261 non-null  float64       
 18  Deadlift2Kg       198030 non-null  float64       
 19  Deadlift3Kg       193530 non-null  float64       
 20  Deadlift4Kg       0 non-null       float64       
 21  Best3DeadliftKg   228623 non-null  float64       
 22  TotalKg           224401 non-null  float64       
 23  Place             233172 non-null  object        
 24  Dots              224373 non-null  float64       
 25  Wilks             224373 non-null  float64       
 26  Glossbrenner      224373 non-null  float64       
 27  Goodlift          224361 non-null  float64       
 28  Tested            232952 non-null  object        
 29  Country           160846 non-null  object        
 30  State             90390 non-null   object        
 31  Federation        233172 non-null  object        
 32  ParentFederation  233172 non-null  object        
 33  Date              233172 non-null  datetime64[ns]
 34  Year              233172 non-null  int32         
 35  MeetCountry       233172 non-null  object        
 36  MeetState         124152 non-null  object        
 37  MeetTown          156834 non-null  object        
 38  MeetName          233172 non-null  object        
dtypes: datetime64[ns](1), float64(21), int32(1), object(16)
memory usage: 68.5+ MB

Step 5: Clean competition results. This will include the squat, bench and deadlift columns¶

A lifter's powerlifting meet is calculated as such (Taken from pg3 of the IPF Rules Book):¶

Each competitor is allowed three attempts on each lift. The lifter’s best valid attempt on each lift counts toward his competition total. If two or more lifters achieve the same total, the lighter lifter ranks above the heavier lifter.¶

From the data, each lifter will have three attempts of each movement. Noted as Squat1kg for first attempt, and Squat3Kg for 3rd attempts etc. Howeover, Squat4Kg, Bench4Kg and Deadlift4Kg all contains null values. This is because lifters may take a 4th attempt on a world record lift. I can remove those columns¶

In [31]:
df2 = df2.drop(['Squat4Kg','Bench4Kg','Deadlift4Kg'], axis=1)
In [32]:
df2.head()
Out[32]:
Name Sex Event Equipment AgeDivision BodyweightKg WeightClassKg Squat1Kg Squat2Kg Squat3Kg Best3SquatKg Bench1Kg Bench2Kg Bench3Kg Best3BenchKg Deadlift1Kg Deadlift2Kg Deadlift3Kg Best3DeadliftKg TotalKg Place Dots Wilks Glossbrenner Goodlift Tested Country State Federation ParentFederation Date Year MeetCountry MeetState MeetTown MeetName
0 Taran Dhanju F SBD Raw Master 1 63.0 63 NaN NaN NaN 90.0 NaN NaN NaN 40.0 NaN NaN NaN 142.5 272.5 1 293.08 292.65 258.42 59.62 Yes Canada ON CPU IPF 2018-06-09 2018 Canada ON Niagara Niagara Open
1 Bryan Sledge M SBD Raw Junior 81.1 83 110.0 125.0 132.5 132.5 -100.0 100.0 NaN 100.0 175.0 195.0 205.0 205.0 437.5 1 299.28 296.15 285.17 61.28 Yes USA AK USAPL IPF 2018-02-24 2018 USA NE NaN Capital City Clash
2 Robert Stolley M SBD Raw Opens 212.0 120+ -172.5 172.5 190.0 190.0 -105.0 -105.0 -105.0 NaN -215.0 -215.0 -215.0 NaN NaN DQ NaN NaN NaN NaN Yes USA NE USAPL IPF 2018-02-24 2018 USA NE NaN Capital City Clash
3 Patrick Mcconnell M SBD Raw Junior 125.5 120+ 150.0 165.0 -177.5 165.0 115.0 125.0 127.5 127.5 225.0 242.5 252.5 252.5 545.0 2 308.49 310.32 296.85 62.14 Yes USA NE USAPL IPF 2018-02-24 2018 USA NE NaN Capital City Clash
4 Mike Kracl M SBD Raw Opens 126.8 120+ 247.5 262.5 272.5 272.5 162.5 170.0 -177.5 170.0 257.5 272.5 285.0 285.0 727.5 1 410.47 413.39 395.42 82.59 Yes USA NE USAPL IPF 2018-02-24 2018 USA NE NaN Capital City Clash
In [33]:
# Check null Values for competition numbers
df2.isna().sum()
Out[33]:
Name                     0
Sex                      0
Event                    0
Equipment                0
AgeDivision              0
BodyweightKg           470
WeightClassKg         1718
Squat1Kg             32095
Squat2Kg             33456
Squat3Kg             36753
Best3SquatKg          3985
Bench1Kg             32520
Bench2Kg             33714
Bench3Kg             37684
Best3BenchKg          4786
Deadlift1Kg          32911
Deadlift2Kg          35142
Deadlift3Kg          39642
Best3DeadliftKg       4549
TotalKg               8771
Place                    0
Dots                  8799
Wilks                 8799
Glossbrenner          8799
Goodlift              8811
Tested                 220
Country              72326
State               142782
Federation               0
ParentFederation         0
Date                     0
Year                     0
MeetCountry              0
MeetState           109020
MeetTown             76338
MeetName                 0
dtype: int64
In [34]:
df2.isna().sum()
Out[34]:
Name                     0
Sex                      0
Event                    0
Equipment                0
AgeDivision              0
BodyweightKg           470
WeightClassKg         1718
Squat1Kg             32095
Squat2Kg             33456
Squat3Kg             36753
Best3SquatKg          3985
Bench1Kg             32520
Bench2Kg             33714
Bench3Kg             37684
Best3BenchKg          4786
Deadlift1Kg          32911
Deadlift2Kg          35142
Deadlift3Kg          39642
Best3DeadliftKg       4549
TotalKg               8771
Place                    0
Dots                  8799
Wilks                 8799
Glossbrenner          8799
Goodlift              8811
Tested                 220
Country              72326
State               142782
Federation               0
ParentFederation         0
Date                     0
Year                     0
MeetCountry              0
MeetState           109020
MeetTown             76338
MeetName                 0
dtype: int64
In [35]:
# Check reason for TotalKg null values
df2[df2['TotalKg'].isnull()]
Out[35]:
Name Sex Event Equipment AgeDivision BodyweightKg WeightClassKg Squat1Kg Squat2Kg Squat3Kg Best3SquatKg Bench1Kg Bench2Kg Bench3Kg Best3BenchKg Deadlift1Kg Deadlift2Kg Deadlift3Kg Best3DeadliftKg TotalKg Place Dots Wilks Glossbrenner Goodlift Tested Country State Federation ParentFederation Date Year MeetCountry MeetState MeetTown MeetName
2 Robert Stolley M SBD Raw Opens 212.00 120+ -172.5 172.5 190.0 190.0 -105.0 -105.0 -105.0 NaN -215.0 -215.0 -215.0 NaN NaN DQ NaN NaN NaN NaN Yes USA NE USAPL IPF 2018-02-24 2018 USA NE NaN Capital City Clash
18 Josh Davis M SBD Raw Junior 58.50 59 -132.5 -132.5 -132.5 NaN -100.0 -105.0 -105.0 NaN 145.0 160.0 -177.5 160.0 NaN DQ NaN NaN NaN NaN Yes USA NE USAPL IPF 2018-02-24 2018 USA NE NaN Capital City Clash
63 Terri Whyte F SBD Raw Opens 163.29 84+ NaN NaN NaN NaN 52.2 -59.0 -59.0 52.2 74.8 81.7 93.0 93.0 NaN DQ NaN NaN NaN NaN Yes USA AK USAPL IPF 2018-12-08 2018 USA AK NaN 18th Annual Ketchikan Winter Meet
74 Kamila Devletova F SBD Raw Sub-Junior 63.00 63 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN DQ NaN NaN NaN NaN Yes NaN NaN FPR IPF 2018-04-18 2018 Russia TUL Tula Russian Sports Schools Classic Powerlifting Ch...
75 Anastasiya Denisova F SBD Raw Sub-Junior 62.15 63 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN DQ NaN NaN NaN NaN Yes Russia NaN FPR IPF 2018-04-18 2018 Russia TUL Tula Russian Sports Schools Classic Powerlifting Ch...
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
233088 Mathys Payet M SBD Raw Junior 81.20 83 200.0 205.0 -210.0 205.0 -120.0 -120.0 -120.0 NaN -220.0 NaN NaN NaN NaN DQ NaN NaN NaN NaN Yes NaN NaN FFForce IPF 2024-01-27 2024 France NaN Saint-Joseph Open du Sud
233090 Mélino Corre M SBD Raw Junior 73.40 74 165.0 -175.0 -175.0 165.0 -115.0 -120.0 -120.0 NaN -170.0 NaN NaN NaN NaN DQ NaN NaN NaN NaN Yes NaN NaN FFForce IPF 2024-01-27 2024 France NaN Saint-Joseph Open du Sud
233126 Lohann Bouda Cambresy M SBD Raw Sub-Junior 63.45 66 -125.0 -135.0 -135.0 NaN 85.0 90.0 -92.5 90.0 155.0 -170.0 -180.0 155.0 NaN DQ NaN NaN NaN NaN Yes NaN NaN FFForce IPF 2024-01-20 2024 France NaN Woippy Championnat Régional Hommes Sub-Juniors/Senior...
233146 Antoine Simon M SBD Raw Opens 79.39 83 187.5 195.0 -200.0 195.0 125.0 130.0 -132.5 130.0 -235.0 NaN NaN NaN NaN DQ NaN NaN NaN NaN Yes NaN NaN FFForce IPF 2024-02-18 2024 France NaN Le Trait Championnat Régional Masters et Seniors FA/PL ...
233167 Mike Maione M SBD Raw Master 2 70.24 74 -125.0 125.0 -130.0 125.0 80.0 85.0 -87.5 85.0 NaN NaN NaN NaN NaN DQ NaN NaN NaN NaN Yes Canada AB CPU IPF 2024-01-20 2024 Canada AB Edmonton Grit Power Open

8771 rows × 36 columns

From the data, I can derive that individuals who has a null value for TotalKg seems to have 'bombed out' from the meet. "Bombing out" in powerlifting refers to the situation where a competitor fails to record a successful lift in one or more of the three competition lifts: squat, bench press, and deadlift. If a lifter fails to complete a successful lift in any of the attempts for a particular lift, they receive no score for that lift, and if they fail to complete any successful lifts in all three lifts, they are said to have "bombed out" of the competition. This means they are unable to post a total and are disqualified from the competition standings. These numbers are denoted with a - infront of the competition numbers¶

In [36]:
# Change TotalKg null values to 0
df2['TotalKg'].fillna(0, inplace=True)

Additionally, I can remove columns of Squat1-3Kg, Bench1-3Kg and Deadlift1-3Kg as the data has provided with 'Best3____Kg' to determine the lifters' best lifts. These columns are not important because our EDA will only include the lifters best performance.¶

In [37]:
df2 = df2.drop(['Squat1Kg','Squat2Kg','Squat3Kg','Bench1Kg','Bench2Kg','Bench3Kg','Deadlift1Kg','Deadlift2Kg','Deadlift3Kg'], axis=1)
In [38]:
# Change Best3SquatKg, Best3BenchKg, Best3DeadliftKg null values to 0
df2 = df2.fillna({'Best3SquatKg': 0, 'Best3BenchKg': 0, 'Best3DeadliftKg': 0})
In [39]:
df2.isna().sum()
Out[39]:
Name                     0
Sex                      0
Event                    0
Equipment                0
AgeDivision              0
BodyweightKg           470
WeightClassKg         1718
Best3SquatKg             0
Best3BenchKg             0
Best3DeadliftKg          0
TotalKg                  0
Place                    0
Dots                  8799
Wilks                 8799
Glossbrenner          8799
Goodlift              8811
Tested                 220
Country              72326
State               142782
Federation               0
ParentFederation         0
Date                     0
Year                     0
MeetCountry              0
MeetState           109020
MeetTown             76338
MeetName                 0
dtype: int64
In [40]:
df2.head()
Out[40]:
Name Sex Event Equipment AgeDivision BodyweightKg WeightClassKg Best3SquatKg Best3BenchKg Best3DeadliftKg TotalKg Place Dots Wilks Glossbrenner Goodlift Tested Country State Federation ParentFederation Date Year MeetCountry MeetState MeetTown MeetName
0 Taran Dhanju F SBD Raw Master 1 63.0 63 90.0 40.0 142.5 272.5 1 293.08 292.65 258.42 59.62 Yes Canada ON CPU IPF 2018-06-09 2018 Canada ON Niagara Niagara Open
1 Bryan Sledge M SBD Raw Junior 81.1 83 132.5 100.0 205.0 437.5 1 299.28 296.15 285.17 61.28 Yes USA AK USAPL IPF 2018-02-24 2018 USA NE NaN Capital City Clash
2 Robert Stolley M SBD Raw Opens 212.0 120+ 190.0 0.0 0.0 0.0 DQ NaN NaN NaN NaN Yes USA NE USAPL IPF 2018-02-24 2018 USA NE NaN Capital City Clash
3 Patrick Mcconnell M SBD Raw Junior 125.5 120+ 165.0 127.5 252.5 545.0 2 308.49 310.32 296.85 62.14 Yes USA NE USAPL IPF 2018-02-24 2018 USA NE NaN Capital City Clash
4 Mike Kracl M SBD Raw Opens 126.8 120+ 272.5 170.0 285.0 727.5 1 410.47 413.39 395.42 82.59 Yes USA NE USAPL IPF 2018-02-24 2018 USA NE NaN Capital City Clash

Let's check whether there are negative values for Best3SquatKg, Best3BenchKg, Best3DeadliftKg, as I assume some federations might have inputted the lifters failed attempt number into the dataset.¶

In [41]:
df2['Best3SquatKg'].min()
Out[41]:
-300.0

It would seem that a negative value for Best3----Kg indicates that a lifter failed all three of the attempts for that specific movement. As TotalKg will indicate '0' if a lifter fails a lift, I can just change negative values from Best3----Kg to 0.¶

In [42]:
def replace_negative(x):
    if x < 0:
        return 0
    else:
        return x
    
df2['Best3SquatKg'] = df2['Best3SquatKg'].apply(replace_negative)
df2['Best3BenchKg'] = df2['Best3BenchKg'].apply(replace_negative)
df2['Best3DeadliftKg'] = df2['Best3DeadliftKg'].apply(replace_negative)

I will now clean the data for Placements and Dots calculation¶

In [43]:
df2['Place'].unique()
Out[43]:
array(['1', 'DQ', '2', '3', '4', '5', '7', '6', 'G', 'NS', '9', '8', '14',
       '13', '12', '11', '10', '17', '16', '15', '18', '19', '26', '25',
       '24', '23', '20', '21', '22', 'DD', '27', '28', '34', '33', '32',
       '31', '30', '29', '40', '39', '38', '37', '36', '35', '63', '62',
       '61', '60', '59', '58', '57', '56', '55', '54', '64', '53', '51',
       '50', '49', '48', '47', '46', '45', '44', '43', '42', '52', '41',
       '65', '67', '66', '79', '77', '76', '75', '74', '73', '72', '71',
       '70', '69', '68', '78', '99', '98', '97', '96', '95', '94', '93',
       '92', '91', '90', '89', '88', '87', '86', '85', '84', '83', '82',
       '81', '80', '113', '112', '111', '110', '109', '108', '107', '106',
       '105', '104', '114', '103', '101', '100', '102', '115', '116',
       '117', '120', '119', '118'], dtype=object)
In [44]:
# Check what are G and NS
df2[df2['Place'] == 'G'].head(5)
Out[44]:
Name Sex Event Equipment AgeDivision BodyweightKg WeightClassKg Best3SquatKg Best3BenchKg Best3DeadliftKg TotalKg Place Dots Wilks Glossbrenner Goodlift Tested Country State Federation ParentFederation Date Year MeetCountry MeetState MeetTown MeetName
94 Rebecca Neal F SBD Raw Opens 69.20 72 97.5 62.5 132.5 292.5 G 297.71 293.32 258.63 60.50 Yes UK NaN VGPF IPF 2018-09-22 2018 Belgium NaN Merchtem Vlaams Kampioenschap - Powerlifting & Bench Pr...
402 Jessica Roberts F SBD Raw Opens 103.32 84+ 135.0 70.0 137.5 342.5 G 288.66 282.60 241.74 60.57 Yes USA NY USAPL IPF 2018-06-03 2018 USA RI NaN Massachusetts And Rhode Island State Championship
403 Jennifer Coen F SBD Raw Junior 78.58 84 112.5 62.5 145.0 320.0 G 304.24 295.79 259.65 62.13 Yes USA NH USAPL IPF 2018-06-03 2018 USA RI NaN Massachusetts And Rhode Island State Championship
404 Emily Claffey F SBD Raw Opens 69.60 72 117.5 75.0 130.0 322.5 G 327.18 322.11 283.97 66.49 Yes USA CT USAPL IPF 2018-06-03 2018 USA RI NaN Massachusetts And Rhode Island State Championship
405 Kelly Wetmore F SBD Raw Junior 67.88 72 122.5 57.5 142.5 322.5 G 331.87 327.83 289.16 67.43 Yes USA CT USAPL IPF 2018-06-03 2018 USA RI NaN Massachusetts And Rhode Island State Championship
In [45]:
df2[df2['Place'] == 'NS'].head(5)
Out[45]:
Name Sex Event Equipment AgeDivision BodyweightKg WeightClassKg Best3SquatKg Best3BenchKg Best3DeadliftKg TotalKg Place Dots Wilks Glossbrenner Goodlift Tested Country State Federation ParentFederation Date Year MeetCountry MeetState MeetTown MeetName
257 枝 和輝 M SBD Raw Opens 93.0 93 0.0 0.0 0.0 0.0 NS NaN NaN NaN NaN Yes NaN NaN JPA IPF 2018-10-14 2018 Japan NaN Tochigi Japanese Businessmen's Powerlifting Championships
286 Hidetoshi Marumoto M SBD Raw Master 4 59.0 59 0.0 0.0 0.0 0.0 NS NaN NaN NaN NaN Yes Japan NaN JPA IPF 2018-10-14 2018 Japan NaN Tochigi Japanese Businessmen's Powerlifting Championships
785 宮城 裕史 M SBD Raw Opens 105.0 105 0.0 0.0 0.0 0.0 NS NaN NaN NaN NaN Yes NaN NaN JPA IPF 2018-04-29 2018 Japan NaN NaN Fukuoka Prefecture Championships
1093 仲田 廣人 M SBD Raw Opens 83.0 83 0.0 0.0 0.0 0.0 NS NaN NaN NaN NaN Yes NaN NaN JPA IPF 2018-02-10 2018 Japan NaN Okinawa Japanese Classic Powerlifting Championships
1094 瀬藤 宗仁 M SBD Raw Opens 83.0 83 0.0 0.0 0.0 0.0 NS NaN NaN NaN NaN Yes NaN NaN JPA IPF 2018-02-10 2018 Japan NaN Okinawa Japanese Classic Powerlifting Championships

It would seem that 'G' are Guest Lifters (the lifter succeeded, but wasn't eligible for awards) and NS for No-Show (the lifter did not show up on the meet day). I can combine DQ (disqualified) and NS as they both represent cases where the lifter did not successfully complete the competition lifts or did not participate in the competition as expected. This can help simplify the analysis and interpretation of the data.¶

In [46]:
df2['Place'] = df2['Place'].replace({'DQ': 'DQ/NS', 'NS': 'DQ/NS'})
In [47]:
df2[df2['Place'] == 'DQ/NS'].head(5)
Out[47]:
Name Sex Event Equipment AgeDivision BodyweightKg WeightClassKg Best3SquatKg Best3BenchKg Best3DeadliftKg TotalKg Place Dots Wilks Glossbrenner Goodlift Tested Country State Federation ParentFederation Date Year MeetCountry MeetState MeetTown MeetName
2 Robert Stolley M SBD Raw Opens 212.00 120+ 190.0 0.0 0.0 0.0 DQ/NS NaN NaN NaN NaN Yes USA NE USAPL IPF 2018-02-24 2018 USA NE NaN Capital City Clash
18 Josh Davis M SBD Raw Junior 58.50 59 0.0 0.0 160.0 0.0 DQ/NS NaN NaN NaN NaN Yes USA NE USAPL IPF 2018-02-24 2018 USA NE NaN Capital City Clash
63 Terri Whyte F SBD Raw Opens 163.29 84+ 0.0 52.2 93.0 0.0 DQ/NS NaN NaN NaN NaN Yes USA AK USAPL IPF 2018-12-08 2018 USA AK NaN 18th Annual Ketchikan Winter Meet
74 Kamila Devletova F SBD Raw Sub-Junior 63.00 63 0.0 0.0 0.0 0.0 DQ/NS NaN NaN NaN NaN Yes NaN NaN FPR IPF 2018-04-18 2018 Russia TUL Tula Russian Sports Schools Classic Powerlifting Ch...
75 Anastasiya Denisova F SBD Raw Sub-Junior 62.15 63 0.0 0.0 0.0 0.0 DQ/NS NaN NaN NaN NaN Yes Russia NaN FPR IPF 2018-04-18 2018 Russia TUL Tula Russian Sports Schools Classic Powerlifting Ch...
In [48]:
df2.isna().sum()
Out[48]:
Name                     0
Sex                      0
Event                    0
Equipment                0
AgeDivision              0
BodyweightKg           470
WeightClassKg         1718
Best3SquatKg             0
Best3BenchKg             0
Best3DeadliftKg          0
TotalKg                  0
Place                    0
Dots                  8799
Wilks                 8799
Glossbrenner          8799
Goodlift              8811
Tested                 220
Country              72326
State               142782
Federation               0
ParentFederation         0
Date                     0
Year                     0
MeetCountry              0
MeetState           109020
MeetTown             76338
MeetName                 0
dtype: int64

In powerlifting, Dots, Wilks, Glossbrenner, and Goodlift refer to different formulas or coefficients used to calculate and compare the strength of lifters across different weight classes.¶

 

  • ##### Dots is very similar to (and drop-in compatible with) the original Wilks formula. It uses an updated, simpler polynomial and is built against data from drug-tested Raw lifters, as opposed to against data from drug-tested Single-ply lifters. The Dots formula was created by Tim Konertz of the BVDK in 2019.  
  • ##### Wilks is the most common formula used for determining Best Lifter in a powerlifting meet.  
  • ##### Glossbrenner was created by Herb Glossbrenner as an update of the Wilks formula. It is most commonly used by GPC-affiliated federations.  
  • ##### Goodlift: IPF GL Points. The successor to IPF Points (2019-01-01 through 2020-04-30). PF GL Points roughly express relative performance to the expected performance of that weight class at an IPF World Championship event, as a percentage.   #### All these formulas and coefficients serve the same purpose: to provide a standardized way of comparing the strength and performance of powerlifters across different weight classes. They allow for a more fair and objective evaluation of lifters' achievements, taking into account the inherent advantages or disadvantages of competing in a particular weight class.
In [49]:
df2[df2['Tested'] == 'No']
Out[49]:
Name Sex Event Equipment AgeDivision BodyweightKg WeightClassKg Best3SquatKg Best3BenchKg Best3DeadliftKg TotalKg Place Dots Wilks Glossbrenner Goodlift Tested Country State Federation ParentFederation Date Year MeetCountry MeetState MeetTown MeetName
In [50]:
df2[df2['Tested'].isna()].head(5)
Out[50]:
Name Sex Event Equipment AgeDivision BodyweightKg WeightClassKg Best3SquatKg Best3BenchKg Best3DeadliftKg TotalKg Place Dots Wilks Glossbrenner Goodlift Tested Country State Federation ParentFederation Date Year MeetCountry MeetState MeetTown MeetName
2728 Nicholas Larson #1 M SBD Raw Junior 94.1 100 185.0 110.0 205.0 500.0 8 316.35 312.37 298.84 65.04 NaN NaN NaN SwissPL IPF 2018-11-27 2018 Switzerland NaN Lausanne Schweizermeisterschaft
2729 Alain Jost M SBD Raw Junior 64.8 67.5 167.5 122.5 205.0 495.0 9 393.21 394.64 383.82 77.92 NaN Switzerland NaN SwissPL IPF 2018-11-27 2018 Switzerland NaN Lausanne Schweizermeisterschaft
2730 Gus Delhees M SBD Raw Junior 90.4 100 0.0 0.0 0.0 0.0 DQ/NS NaN NaN NaN NaN NaN NaN NaN SwissPL IPF 2018-11-27 2018 Switzerland NaN Lausanne Schweizermeisterschaft
2731 Jana Kamberi F SBD Raw Opens 65.9 67.5 147.5 60.0 140.0 347.5 1 363.81 360.89 318.48 73.93 NaN Switzerland NaN SwissPL IPF 2018-11-27 2018 Switzerland NaN Lausanne Schweizermeisterschaft
2732 Andrea Marx F SBD Raw Opens 68.8 75 115.0 72.5 157.5 345.0 2 352.30 347.38 306.32 71.59 NaN Switzerland NaN SwissPL IPF 2018-11-27 2018 Switzerland NaN Lausanne Schweizermeisterschaft

Since most of the data are Tested meets, I will only be using the Dots formula in this EDA¶

In [51]:
df2 = df2.drop(['Wilks','Glossbrenner','Goodlift'], axis=1)
In [52]:
# since those DQ and NS will also include null values in DOTS, lets change them to 0
df2['Dots'].fillna(0, inplace=True)
In [53]:
df2.head()
Out[53]:
Name Sex Event Equipment AgeDivision BodyweightKg WeightClassKg Best3SquatKg Best3BenchKg Best3DeadliftKg TotalKg Place Dots Tested Country State Federation ParentFederation Date Year MeetCountry MeetState MeetTown MeetName
0 Taran Dhanju F SBD Raw Master 1 63.0 63 90.0 40.0 142.5 272.5 1 293.08 Yes Canada ON CPU IPF 2018-06-09 2018 Canada ON Niagara Niagara Open
1 Bryan Sledge M SBD Raw Junior 81.1 83 132.5 100.0 205.0 437.5 1 299.28 Yes USA AK USAPL IPF 2018-02-24 2018 USA NE NaN Capital City Clash
2 Robert Stolley M SBD Raw Opens 212.0 120+ 190.0 0.0 0.0 0.0 DQ/NS 0.00 Yes USA NE USAPL IPF 2018-02-24 2018 USA NE NaN Capital City Clash
3 Patrick Mcconnell M SBD Raw Junior 125.5 120+ 165.0 127.5 252.5 545.0 2 308.49 Yes USA NE USAPL IPF 2018-02-24 2018 USA NE NaN Capital City Clash
4 Mike Kracl M SBD Raw Opens 126.8 120+ 272.5 170.0 285.0 727.5 1 410.47 Yes USA NE USAPL IPF 2018-02-24 2018 USA NE NaN Capital City Clash
In [54]:
df2.shape
Out[54]:
(233172, 24)
In [55]:
# Dropping null values in 'Tested' as I want a fair data comparison of only tested lifters
df2 = df2.dropna(subset=['Tested'])
df2 = df2.drop(['Tested'], axis=1)
In [56]:
df2.isna().sum()
Out[56]:
Name                     0
Sex                      0
Event                    0
Equipment                0
AgeDivision              0
BodyweightKg           467
WeightClassKg         1710
Best3SquatKg             0
Best3BenchKg             0
Best3DeadliftKg          0
TotalKg                  0
Place                    0
Dots                     0
Country              72181
State               142562
Federation               0
ParentFederation         0
Date                     0
Year                     0
MeetCountry              0
MeetState           108800
MeetTown             76338
MeetName                 0
dtype: int64

Step 6: Clean weight classes¶

In [57]:
# Check null values for WeightClassKG
df2[df2['WeightClassKg'].isna()].head(5)
Out[57]:
Name Sex Event Equipment AgeDivision BodyweightKg WeightClassKg Best3SquatKg Best3BenchKg Best3DeadliftKg TotalKg Place Dots Country State Federation ParentFederation Date Year MeetCountry MeetState MeetTown MeetName
10424 Viktor Vasilyev M SBD Raw Master 3 110.00 NaN 170.0 155.0 170.0 495.0 1 293.17 Russia NaN FPR IPF 2018-02-24 2018 Russia KGD Kaliningrad Kaliningrad Oblast Powerlifting Championships
10425 Sergey Lobych M SBD Raw Master 1 81.35 NaN 155.0 120.0 195.0 470.0 3 320.94 Russia NaN FPR IPF 2018-02-24 2018 Russia KGD Kaliningrad Kaliningrad Oblast Powerlifting Championships
10426 Sergey Raschupkin M SBD Raw Master 1 96.85 NaN 200.0 135.0 215.0 550.0 2 343.37 Russia NaN FPR IPF 2018-02-24 2018 Russia KGD Kaliningrad Kaliningrad Oblast Powerlifting Championships
10427 Aleksey Petrov M SBD Raw Master 1 111.00 NaN 205.0 165.0 232.5 602.5 1 355.64 Russia NaN FPR IPF 2018-02-24 2018 Russia KGD Kaliningrad Kaliningrad Oblast Powerlifting Championships
13404 Małgorzata Otto F SBD Raw Master 1 57.80 NaN 110.0 60.0 125.0 295.0 1 334.93 Poland NaN PZKFiTS IPF 2018-03-17 2018 Poland NaN Skierniewice Polish Classic Powerlifting Cup
In [58]:
df2['WeightClassKg'].unique()
Out[58]:
array(['63', '83', '120+', '120', '93', '74', '66', '59', '105', '72',
       '84', '84+', '57', '52', '47', '43', '72+', '53', '93+', '48',
       '40', '57+', '63+', '105+', '35', nan, '83+', '74+', '82.5', '125',
       '75', '110', '100', '90', '90+', '67.5', '60', '56', '125+', '44',
       '52+', '+', '69', '76', '69+', '76+', '39'], dtype=object)

In the IPF rule book, these are the weight categories stated. I will have to manipulate and remove the data of the outliers¶

image.png

In [59]:
# Finding the shape of IPF weight classes
ipf_weight_classes = [
    '53','59','66','74','83','93','105','120','120+',
    '47','52','57','63','69','76','84','84+'
]

df2[df2['WeightClassKg'].isin(ipf_weight_classes)].shape
Out[59]:
(220209, 23)

USAPL stands for USA Powerlifting, which is the United States affiliate to the International Powerlifting Federation (IPF). It's the national governing body for the sport of powerlifting in the United States. Eventhough USAPL is an affliate to IPF, most (if not all) of their hosted competitions doesn't follow the IPF's weight classses as they have autonomy in setting their own competition rules.¶

These are the weight class structure provided in their own USAPL rulebook:¶

image.png

In [60]:
# Finding the shape of USAPL weight classes
usapl_weight_classes = [
    '52', '56', '60', '67.5', '75', '82.5', '90', '100', '110', '125', '140', '140+',
    '44', '48', '52', '56', '60', '67.5', '82.5', '90', '100', '100+'
]

df2[df2['WeightClassKg'].isin(usapl_weight_classes)].shape
Out[60]:
(7064, 23)
In [61]:
# Finding the remaining weight classes outliers
remaining_weight_classes = set(df2['WeightClassKg']) - set(ipf_weight_classes) - set(usapl_weight_classes)
print(remaining_weight_classes)
df2[~df2['WeightClassKg'].isin(ipf_weight_classes) & ~df2['WeightClassKg'].isin(usapl_weight_classes) & ~df2['WeightClassKg'].isnull()]

{'74+', '72', '72+', '35', '90+', '40', '93+', '39', nan, '69+', '52+', '43', '63+', '76+', '83+', '125+', '57+', '+', '105+'}
Out[61]:
Name Sex Event Equipment AgeDivision BodyweightKg WeightClassKg Best3SquatKg Best3BenchKg Best3DeadliftKg TotalKg Place Dots Country State Federation ParentFederation Date Year MeetCountry MeetState MeetTown MeetName
41 Brittney Woznicki F SBD Raw Junior 68.10 72 105.0 67.5 122.5 295.0 1 303.01 USA OH USAPL IPF 2018-08-04 2018 USA OH NaN Dog Days of Summer Powerlifting
46 Kenda Conley F SBD Raw Opens 67.13 72 99.8 63.5 129.3 292.6 2 303.04 USA AK USAPL IPF 2018-12-08 2018 USA AK NaN 18th Annual Ketchikan Winter Meet
64 Gale Lindemann F SBD Raw Master 2 64.18 72 52.2 34.0 65.8 152.0 7 161.65 USA AK USAPL IPF 2018-12-08 2018 USA AK NaN 18th Annual Ketchikan Winter Meet
68 Vanessa Head F SBD Raw Master 1 70.76 72 95.3 61.2 115.7 272.2 3 273.61 USA AK USAPL IPF 2018-12-08 2018 USA AK NaN 18th Annual Ketchikan Winter Meet
87 Viktoriya Panova F SBD Raw Sub-Junior 39.60 43 50.0 30.0 75.0 155.0 4 230.14 NaN NaN FPR IPF 2018-04-18 2018 Russia TUL Tula Russian Sports Schools Classic Powerlifting Ch...
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
227740 Elena Galbucci F SBD Raw Junior 42.75 43 77.5 0.0 105.0 0.0 DQ/NS 0.00 NaN NaN FIPL IPF 2023-11-17 2023 Italy NaN S.Zenone al Lambro 3 gara nazionale trofeo next generation classi...
228582 Praveena Thurairajasingam-Mukunthan F SBD Raw Junior 42.00 43 97.5 45.0 122.5 265.0 1 378.44 NaN NaN APU IPF 2023-12-02 2023 Australia WA NaN WA State Classic Powerlifting Championships
230421 Gwen Marsden F SBD Raw Junior 42.60 43 112.5 75.0 142.5 330.0 1 466.05 UK NaN EPA IPF 2024-02-03 2024 England NaN Cambridge Cambridge vs Oxford Powerlifting Varsity
230931 Christine Mechaley F SBD Raw Sub-Junior 42.80 43 45.0 25.0 0.0 0.0 DQ/NS 0.00 USA NE AMP IPF 2024-01-20 2024 USA NE Lincoln Capital City Clash
232895 Mayleen Holder F SBD Raw Sub-Junior 42.25 43 85.0 47.5 100.0 232.5 1 330.49 NaN NaN FFForce IPF 2024-01-27 2024 France NaN Woippy Championnat Régional Femmes FA/PL Grand Est

10629 rows × 23 columns

It appears that a significant portion of the outliers consists of data from youth weight categories and various affiliates worldwide. Therefore, it is essential to exclude these numbers as the focus of this analysis is to adhere to the standardized rules of the International Powerlifting Federation (IPF). This also includes removing data from USAPL competitions too.¶

In [62]:
# Checking null values
df2[df2['WeightClassKg'].isna()]
Out[62]:
Name Sex Event Equipment AgeDivision BodyweightKg WeightClassKg Best3SquatKg Best3BenchKg Best3DeadliftKg TotalKg Place Dots Country State Federation ParentFederation Date Year MeetCountry MeetState MeetTown MeetName
10424 Viktor Vasilyev M SBD Raw Master 3 110.00 NaN 170.0 155.0 170.0 495.0 1 293.17 Russia NaN FPR IPF 2018-02-24 2018 Russia KGD Kaliningrad Kaliningrad Oblast Powerlifting Championships
10425 Sergey Lobych M SBD Raw Master 1 81.35 NaN 155.0 120.0 195.0 470.0 3 320.94 Russia NaN FPR IPF 2018-02-24 2018 Russia KGD Kaliningrad Kaliningrad Oblast Powerlifting Championships
10426 Sergey Raschupkin M SBD Raw Master 1 96.85 NaN 200.0 135.0 215.0 550.0 2 343.37 Russia NaN FPR IPF 2018-02-24 2018 Russia KGD Kaliningrad Kaliningrad Oblast Powerlifting Championships
10427 Aleksey Petrov M SBD Raw Master 1 111.00 NaN 205.0 165.0 232.5 602.5 1 355.64 Russia NaN FPR IPF 2018-02-24 2018 Russia KGD Kaliningrad Kaliningrad Oblast Powerlifting Championships
13404 Małgorzata Otto F SBD Raw Master 1 57.80 NaN 110.0 60.0 125.0 295.0 1 334.93 Poland NaN PZKFiTS IPF 2018-03-17 2018 Poland NaN Skierniewice Polish Classic Powerlifting Cup
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
232196 Jassim Messina M SBD Raw Junior 82.40 NaN 190.0 105.0 240.0 535.0 4 362.65 NaN NaN KDKS IPF 2024-02-03 2024 Switzerland NaN Olten Vereinswettkampf Maschinerie
232197 Fridolin Hauser M SBD Raw Opens 89.00 NaN 185.0 117.5 207.5 510.0 5 331.64 NaN NaN KDKS IPF 2024-02-03 2024 Switzerland NaN Olten Vereinswettkampf Maschinerie
232214 Stefan Bur M SBD Raw Master 2 73.70 NaN 135.0 95.0 140.0 370.0 10 268.46 NaN NaN KDKS IPF 2024-02-03 2024 Switzerland NaN Olten Vereinswettkampf Maschinerie
232601 Josh Dungate M SBD Raw Junior NaN NaN 0.0 0.0 0.0 0.0 DQ/NS 0.00 England SW EPA IPF 2024-02-17 2024 England SW Gloucester Glos Power Development Competition
232791 Simon Parsons M SBD Raw Opens NaN NaN 0.0 0.0 0.0 0.0 DQ/NS 0.00 England NaN EPA IPF 2024-02-04 2024 England NaN London Greater London BUCS Qualifier

1710 rows × 23 columns

Given the nature of the dataset, it would be impossible to manipluate null values, even with the data of the lifter's bodyweight. There are too many variables such as varying weight classes from affliates around the world, and even justifying if a meet is following IPF's/USAPL 's weight category ruling is an impossible task. Therefore, I will exclude null values from the dataset.¶

In [63]:
df2.shape
Out[63]:
(232952, 23)
In [64]:
df2[df2['WeightClassKg'].isin(ipf_weight_classes)].shape
Out[64]:
(220209, 23)

It would appear that only 10000~ are filtered after the justifications made. I will also need to remove any null values from the BodyweightKg column as I don't want any unknown values in the dataset¶

In [65]:
df2[df2['BodyweightKg'].isnull()].shape
Out[65]:
(467, 23)
In [66]:
# Assigning back to df2
df2 = df2[df2['BodyweightKg'].notnull()]
df2 = df2[df2['WeightClassKg'].isin(ipf_weight_classes)]

df2 = df2.reset_index(drop=True)

Check for any outliers for Male/Female¶

In [67]:
df2['Sex'].unique()
Out[67]:
array(['F', 'M', 'Mx'], dtype=object)
In [68]:
df2[df2['Sex'] == 'Mx']
Out[68]:
Name Sex Event Equipment AgeDivision BodyweightKg WeightClassKg Best3SquatKg Best3BenchKg Best3DeadliftKg TotalKg Place Dots Country State Federation ParentFederation Date Year MeetCountry MeetState MeetTown MeetName
110934 Elaine Bradley Mx SBD Raw Opens 88.0 84+ 112.5 57.5 130.0 300.0 1 196.22 USA GA USAPL IPF 2020-11-07 2020 USA GA NaN Viking Strength Sports
In [69]:
df2 = df2[df2['Sex'] != 'Mx']

Let's check if there are outliers, where the Male and Female competitors might have different values than in the IPF rule book¶

In [70]:
male_ipf_weight_classes = [
    '53','59','66','74','83','93','105','120','120+',
]

# Create a boolean mask for 'Sex' is 'M'
male_mask = df2['Sex'] == 'M'

# Check outliers 
df2[male_mask & ~df2['WeightClassKg'].isin(male_ipf_weight_classes)]
Out[70]:
Name Sex Event Equipment AgeDivision BodyweightKg WeightClassKg Best3SquatKg Best3BenchKg Best3DeadliftKg TotalKg Place Dots Country State Federation ParentFederation Date Year MeetCountry MeetState MeetTown MeetName
26759 Justice Dilley M SBD Raw Sub-Junior 51.5 52 0.0 70.0 125.0 0.0 DQ/NS 0.00 USA WI USAPL IPF 2018-03-11 2018 USA WI NaN WHSPA State Championship
26760 Dylan Daly-Schreiner M SBD Raw Sub-Junior 50.9 52 62.5 42.5 107.5 212.5 8 207.74 USA WI USAPL IPF 2018-03-11 2018 USA WI NaN WHSPA State Championship
26761 Carson Junemann M SBD Raw Sub-Junior 51.4 52 77.5 40.0 110.0 227.5 7 220.32 USA WI USAPL IPF 2018-03-11 2018 USA WI NaN WHSPA State Championship
26762 Dawson Jenkin M SBD Raw Sub-Junior 49.9 52 85.0 45.0 102.5 232.5 6 231.74 USA WI USAPL IPF 2018-03-11 2018 USA WI NaN WHSPA State Championship
26763 Jacob Matoushek M SBD Raw Sub-Junior 49.3 52 80.0 50.0 105.0 235.0 5 237.06 USA WI USAPL IPF 2018-03-11 2018 USA WI NaN WHSPA State Championship
26764 Christian Shigouri M SBD Raw Sub-Junior 51.2 52 80.0 55.0 130.0 265.0 4 257.60 USA WI USAPL IPF 2018-03-11 2018 USA WI NaN WHSPA State Championship
26765 Lane Lee M SBD Raw Sub-Junior 52.0 52 95.0 55.0 125.0 275.0 3 263.39 USA WI USAPL IPF 2018-03-11 2018 USA WI NaN WHSPA State Championship
26766 Forest Thoreson M SBD Raw Sub-Junior 51.8 52 95.0 67.5 137.5 300.0 2 288.39 USA WI USAPL IPF 2018-03-11 2018 USA WI NaN WHSPA State Championship
26767 Brian Bauch M SBD Raw Sub-Junior 50.7 52 115.0 67.5 137.5 320.0 1 314.03 USA WI USAPL IPF 2018-03-11 2018 USA WI NaN WHSPA State Championship
26768 Alan Jimenez M SBD Raw Sub-Junior 51.1 52 75.0 40.0 87.5 202.5 9 197.22 USA WI USAPL IPF 2018-03-11 2018 USA WI NaN WHSPA State Championship
In [71]:
female_ipf_weight_classes = [
    '47','52','57','63','69','76','84','84+'
]
# Create a boolean mask where 'Sex' is 'F'
female_mask = df2['Sex'] == 'F'

# Check outliers 
df2[female_mask & ~df2['WeightClassKg'].isin(female_ipf_weight_classes)]
Out[71]:
Name Sex Event Equipment AgeDivision BodyweightKg WeightClassKg Best3SquatKg Best3BenchKg Best3DeadliftKg TotalKg Place Dots Country State Federation ParentFederation Date Year MeetCountry MeetState MeetTown MeetName
117994 Oksana Rimkienė F SBD Raw Master 1 55.75 59 85.0 40.0 105.0 230.0 1 267.39 NaN NaN LJTF IPF 2021-05-29 2021 Lithuania NaN Pravieniškės Classic National Championship
In [72]:
# Create a combine mask to remove these outliers
combined_mask = ((df2['Sex'] == 'M') & (~df2['WeightClassKg'].isin(male_ipf_weight_classes))) | \
                ((df2['Sex'] == 'F') & (~df2['WeightClassKg'].isin(female_ipf_weight_classes)))

df2 = df2[~combined_mask]
In [73]:
df2.shape
Out[73]:
(220141, 23)
In [74]:
df2.head(5)
Out[74]:
Name Sex Event Equipment AgeDivision BodyweightKg WeightClassKg Best3SquatKg Best3BenchKg Best3DeadliftKg TotalKg Place Dots Country State Federation ParentFederation Date Year MeetCountry MeetState MeetTown MeetName
0 Taran Dhanju F SBD Raw Master 1 63.0 63 90.0 40.0 142.5 272.5 1 293.08 Canada ON CPU IPF 2018-06-09 2018 Canada ON Niagara Niagara Open
1 Bryan Sledge M SBD Raw Junior 81.1 83 132.5 100.0 205.0 437.5 1 299.28 USA AK USAPL IPF 2018-02-24 2018 USA NE NaN Capital City Clash
2 Robert Stolley M SBD Raw Opens 212.0 120+ 190.0 0.0 0.0 0.0 DQ/NS 0.00 USA NE USAPL IPF 2018-02-24 2018 USA NE NaN Capital City Clash
3 Patrick Mcconnell M SBD Raw Junior 125.5 120+ 165.0 127.5 252.5 545.0 2 308.49 USA NE USAPL IPF 2018-02-24 2018 USA NE NaN Capital City Clash
4 Mike Kracl M SBD Raw Opens 126.8 120+ 272.5 170.0 285.0 727.5 1 410.47 USA NE USAPL IPF 2018-02-24 2018 USA NE NaN Capital City Clash

7. Clean rest of the columns¶

In [75]:
# Check the remaining null values
df2.isna().sum()
Out[75]:
Name                     0
Sex                      0
Event                    0
Equipment                0
AgeDivision              0
BodyweightKg             0
WeightClassKg            0
Best3SquatKg             0
Best3BenchKg             0
Best3DeadliftKg          0
TotalKg                  0
Place                    0
Dots                     0
Country              68956
State               135907
Federation               0
ParentFederation         0
Date                     0
Year                     0
MeetCountry              0
MeetState           103666
MeetTown             71338
MeetName                 0
dtype: int64
In [76]:
# Checking dataset on null values in 'Country' column
df2[df2['Country'].isnull()].head(5)
Out[76]:
Name Sex Event Equipment AgeDivision BodyweightKg WeightClassKg Best3SquatKg Best3BenchKg Best3DeadliftKg TotalKg Place Dots Country State Federation ParentFederation Date Year MeetCountry MeetState MeetTown MeetName
69 Vera Gorelova F SBD Raw Sub-Junior 81.55 84 130.0 57.5 120.0 307.5 1 287.01 NaN NaN FPR IPF 2018-04-18 2018 Russia TUL Tula Russian Sports Schools Classic Powerlifting Ch...
70 Kamila Devletova F SBD Raw Sub-Junior 63.00 63 0.0 0.0 0.0 0.0 DQ/NS 0.00 NaN NaN FPR IPF 2018-04-18 2018 Russia TUL Tula Russian Sports Schools Classic Powerlifting Ch...
72 Ulyana Ovsyannikova F SBD Raw Sub-Junior 62.35 63 70.0 40.0 85.0 195.0 3 211.06 NaN NaN FPR IPF 2018-04-18 2018 Russia TUL Tula Russian Sports Schools Classic Powerlifting Ch...
73 Kseniya Panteleeva F SBD Raw Sub-Junior 60.65 63 90.0 47.5 90.0 227.5 1 250.49 NaN NaN FPR IPF 2018-04-18 2018 Russia TUL Tula Russian Sports Schools Classic Powerlifting Ch...
77 Karina Trofimova F SBD Raw Sub-Junior 51.35 52 45.0 27.5 60.0 132.5 4 162.93 NaN NaN FPR IPF 2018-04-18 2018 Russia TUL Tula Russian Sports Schools Classic Powerlifting Ch...

'MeetCountry' seems to indicate the country the competition was held whereas 'Country' indicates where the lifter was born. It seems that the lifter's birth country contains null values. I will change those null values to 'Unknown' for future exploratory analysis.¶

In [77]:
# Changing null values to Unknown
df2['Country'].fillna('Unknown', inplace=True)

I also don't need the State, MeetState and MeetTown columns.¶

In [78]:
df2 = df2.drop(['State','MeetState','MeetTown'], axis=1)
In [79]:
df2.head(5)
Out[79]:
Name Sex Event Equipment AgeDivision BodyweightKg WeightClassKg Best3SquatKg Best3BenchKg Best3DeadliftKg TotalKg Place Dots Country Federation ParentFederation Date Year MeetCountry MeetName
0 Taran Dhanju F SBD Raw Master 1 63.0 63 90.0 40.0 142.5 272.5 1 293.08 Canada CPU IPF 2018-06-09 2018 Canada Niagara Open
1 Bryan Sledge M SBD Raw Junior 81.1 83 132.5 100.0 205.0 437.5 1 299.28 USA USAPL IPF 2018-02-24 2018 USA Capital City Clash
2 Robert Stolley M SBD Raw Opens 212.0 120+ 190.0 0.0 0.0 0.0 DQ/NS 0.00 USA USAPL IPF 2018-02-24 2018 USA Capital City Clash
3 Patrick Mcconnell M SBD Raw Junior 125.5 120+ 165.0 127.5 252.5 545.0 2 308.49 USA USAPL IPF 2018-02-24 2018 USA Capital City Clash
4 Mike Kracl M SBD Raw Opens 126.8 120+ 272.5 170.0 285.0 727.5 1 410.47 USA USAPL IPF 2018-02-24 2018 USA Capital City Clash
In [80]:
df2.isna().sum()
Out[80]:
Name                0
Sex                 0
Event               0
Equipment           0
AgeDivision         0
BodyweightKg        0
WeightClassKg       0
Best3SquatKg        0
Best3BenchKg        0
Best3DeadliftKg     0
TotalKg             0
Place               0
Dots                0
Country             0
Federation          0
ParentFederation    0
Date                0
Year                0
MeetCountry         0
MeetName            0
dtype: int64
In [81]:
# Dropping ParentFederation and Equipment as it is all IPF, and Raw lifters
df2 = df2.drop(['ParentFederation','Equipment'], axis=1)

Perfect! Finally I will look through any duplicated values in the dataset.¶

In [82]:
df2.shape
Out[82]:
(220141, 18)
In [83]:
# Check all exact values of rows
df2[df2.duplicated(keep=False)]
Out[83]:
Name Sex Event AgeDivision BodyweightKg WeightClassKg Best3SquatKg Best3BenchKg Best3DeadliftKg TotalKg Place Dots Country Federation Date Year MeetCountry MeetName
66 Hannah Mathie F SBD Junior 56.80 57 142.5 65.0 142.5 350.0 1 401.92 USA USAPL 2018-08-04 2018 USA Dog Days of Summer Powerlifting
68 Hannah Mathie F SBD Junior 56.80 57 142.5 65.0 142.5 350.0 1 401.92 USA USAPL 2018-08-04 2018 USA Dog Days of Summer Powerlifting
141 Russel Nguyen M SBD Junior 76.70 83 205.0 135.0 240.0 580.0 1 410.27 USA USAPL 2018-05-19 2018 USA Long Island Spring Showdown
144 Amy Zhu F SBD Junior 56.50 57 112.5 60.0 117.5 290.0 1 334.18 USA USAPL 2018-05-19 2018 USA Long Island Spring Showdown
145 Amy Zhu F SBD Junior 56.50 57 112.5 60.0 117.5 290.0 1 334.18 USA USAPL 2018-05-19 2018 USA Long Island Spring Showdown
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
219743 Jessica Marshall #1 F SBD Master 3 63.40 69 117.5 60.0 152.5 330.0 1 353.56 USA AMP 2024-02-03 2024 USA DMV Championships
219766 Michelle Harrell F SBD Master 1 104.90 84+ 150.0 95.0 175.0 420.0 1 352.01 USA AMP 2024-02-17 2024 USA Lake Pleasant Classic
219771 Michelle Harrell F SBD Master 1 104.90 84+ 150.0 95.0 175.0 420.0 1 352.01 USA AMP 2024-02-17 2024 USA Lake Pleasant Classic
219778 Cherie Tan F SBD Master 1 67.97 69 80.0 35.0 87.5 202.5 1 208.23 USA AMP 2024-01-20 2024 USA 7th Annual Jacksonville Open Championship
220041 Cherie Tan F SBD Master 1 67.97 69 80.0 35.0 87.5 202.5 1 208.23 USA AMP 2024-01-20 2024 USA 7th Annual Jacksonville Open Championship

6797 rows × 18 columns

In [84]:
# Drop duplicated values
df2 = df2.drop_duplicates(keep='first')
In [85]:
df2[df2.duplicated(keep=False)]
Out[85]:
Name Sex Event AgeDivision BodyweightKg WeightClassKg Best3SquatKg Best3BenchKg Best3DeadliftKg TotalKg Place Dots Country Federation Date Year MeetCountry MeetName
In [86]:
# Renaming of columns
df3 = df2.rename(columns = {
    'Name': 'name',
    'Sex': 'sex',
    'Event': 'event',
    'AgeDivision': 'age_division',
    'BodyweightKg': 'bodyweight',
    'WeightClassKg': 'weight_class',
    'Best3SquatKg': 'best_squat',
    'Best3BenchKg': 'best_bench',
    'Best3DeadliftKg': 'best_deadlift',
    'TotalKg': 'total',
    'Place': 'place',
    'Dots': 'dots_score',
    'Country': 'birth_country',
    'Federation': 'federation',
    'Date': 'meet_date',
    'Year': 'meet_year',
    'MeetCountry': 'meet_country',
    'MeetName': 'meet_name'
                })

# Remove hhmmss from date
df3['meet_date'] = df3['meet_date'].dt.date

# Sort values with meet_date
df3 = df3.sort_values(by='meet_date')
df3 = df3.reset_index(drop=True)
In [87]:
df3.head(10)
Out[87]:
name sex event age_division bodyweight weight_class best_squat best_bench best_deadlift total place dots_score birth_country federation meet_date meet_year meet_country meet_name
0 Daniel Gonzalez M SBD Opens 93.7 105 185.0 115.0 220.0 520.0 3 329.67 Unknown FECHIPO 2018-01-01 2018 Chile Nacional Chile
1 Miguel Franco Gallado M SBD Opens 77.1 83 155.0 105.0 200.0 460.0 5 324.34 Unknown FECHIPO 2018-01-01 2018 Chile Nacional Chile
2 Sebastian Camilo Troncozo Gonzalez M SBD Opens 78.6 83 167.5 100.0 192.5 460.0 6 320.55 Unknown FECHIPO 2018-01-01 2018 Chile Nacional Chile
3 Nicolás Andrés Rojas Grandón M SBD Opens 81.0 83 170.0 112.5 185.0 467.5 4 320.03 Unknown FECHIPO 2018-01-01 2018 Chile Nacional Chile
4 Francisco Andrés Peña Hermosilla M SBD Opens 87.3 93 170.0 105.0 200.0 475.0 5 311.98 Unknown FECHIPO 2018-01-01 2018 Chile Nacional Chile
5 Astrid Carolina Kroff Rosentreter F SBD Opens 60.0 63 80.0 45.0 120.0 245.0 2 271.59 Unknown FECHIPO 2018-01-01 2018 Chile Nacional Chile
6 Javier Ignacio Ventura Lovera M SBD Junior 70.7 74 132.5 100.0 165.0 397.5 4 296.54 Unknown FECHIPO 2018-01-01 2018 Chile Nacional Chile
7 Francisco Villarroel Machuca M SBD Junior 72.7 74 162.5 105.0 192.5 460.0 2 336.77 Unknown FECHIPO 2018-01-01 2018 Chile Nacional Chile
8 Paolo Araya M SBD Opens 90.9 93 200.0 125.0 220.0 545.0 2 350.65 Unknown FECHIPO 2018-01-01 2018 Chile Nacional Chile
9 Anyelo Riquelme Urra M SBD Junior 69.7 74 170.0 130.0 220.0 520.0 1 391.77 Unknown FECHIPO 2018-01-01 2018 Chile Nacional Chile

Finally, this dataset is cleaned and ready to be used for EDA¶

Exploratory Data Analysis¶

Let's check out the most popular weight classes!¶

In [88]:
fig, axes = plt.subplots(1, 2, figsize=(12, 6))

sort_df = df3.sort_values(by='weight_class', ascending=True)

sns.histplot(data=sort_df[sort_df['sex'] == 'M'], x='weight_class', ax=axes[0], color='skyblue')
axes[0].set_title('Male Weight Classes')

sns.histplot(data=sort_df[sort_df['sex'] == 'F'], x='weight_class', ax=axes[1], color='lightcoral')
axes[1].set_title('Female Weight Classes')

plt.tight_layout()
plt.show()

It would appear that the most popular male weight classes are U83, and U93. While U63 is the most popular for females. I wonder what is the distribution between male and female lifters are!¶

In [89]:
gender_counts = df3['sex'].value_counts()

plt.figure(figsize=(8, 4))
plt.pie(gender_counts, labels=gender_counts.index, autopct='%1.1f%%', colors=['skyblue', 'lightcoral'])
plt.title('Male vs Female Lifters')
plt.axis('equal')  # Equal aspect ratio so that pie is drawn as a circle
plt.show()

As expected, there are significantly higher male lifters than their counterpart! That being said, I wonder who is the strongest male and female. I will use DOTS score to determine this question, as it is a ratio between a lifter's body weight and the weight they're lifting.¶

In [90]:
df3.iloc[df3[df3['sex'] == 'M']['dots_score'].idxmax()]
Out[90]:
name             Taylor Atwood
sex                          M
event                      SBD
age_division             Opens
bodyweight               73.63
weight_class                74
best_squat               303.0
best_bench               195.0
best_deadlift            340.5
total                    838.5
place                        1
dots_score              608.77
birth_country              USA
federation               USAPL
meet_date           2021-06-14
meet_year                 2021
meet_country               USA
meet_name        Raw Nationals
Name: 124790, dtype: object
In [91]:
df3.iloc[df3[df3['sex'] == 'F']['dots_score'].idxmax()]
Out[91]:
name                  Prescillia Bavoil
sex                                   F
event                               SBD
age_division                      Opens
bodyweight                         62.6
weight_class                         63
best_squat                        213.0
best_bench                        112.5
best_deadlift                     231.0
total                             556.5
place                                 1
dots_score                       600.86
birth_country                    France
federation                          IPF
meet_date                    2022-09-23
meet_year                          2022
meet_country                    England
meet_name        Arnold Sports Festival
Name: 164043, dtype: object

Taylor Atwood and Lya (Prescillia Bavoil) are widely recognized figures in the powerlifting community, boasting over a decade of experience each in this demanding sport. It wasn't surprising to see their names appearing! Now I'm inclinced to know who lifted the most weight in a competition.¶

In [92]:
df3.iloc[df3[df3['sex'] == 'M']['total'].idxmax()]
Out[92]:
name                     Ray Williams #1
sex                                    M
event                                SBD
age_division                       Opens
bodyweight                         190.4
weight_class                        120+
best_squat                         490.0
best_bench                         247.5
best_deadlift                      375.0
total                             1112.5
place                                  1
dots_score                        562.77
birth_country                        USA
federation                         USAPL
meet_date                     2019-03-02
meet_year                           2019
meet_country                         USA
meet_name        Arnold SBD Pro American
Name: 50656, dtype: object

RAY WILLIAMS LIFTED 1112.5 KGS???? Looking through wikipedia:¶

image.png

Ray Williams doesn't just lift weights, he lifts American Bisons! If that's not superhuman, I don't know what is..¶

image-2.png

Now back to your average lifter (notice the huge difference in size 🤣): In 2022, I competed as a U59 lifter, indicating that my weight class was under 59kg. During these competitions, my bodyweight hovered around the 58-59kg mark, aligning closely with the requirements of the ipf weight class. Though not as strong as Ray Williams, I did managed to achieve decent results in both meets!¶

image.png

Let's take a look at my numbers!¶

In [93]:
df3.loc[df3['name'] == 'Muhammad Ghazi Bin Hishamuddin', 'name']
Out[93]:
160540    Muhammad Ghazi Bin Hishamuddin
172553    Muhammad Ghazi Bin Hishamuddin
Name: name, dtype: object
In [94]:
# 160552
df3.iloc[160540]
Out[94]:
name             Muhammad Ghazi Bin Hishamuddin
sex                                           M
event                                       SBD
age_division                              Opens
bodyweight                                58.25
weight_class                                 59
best_squat                                150.0
best_bench                                105.0
best_deadlift                             187.5
total                                     442.5
place                                         2
dots_score                               382.82
birth_country                           Unknown
federation                                   PS
meet_date                            2022-08-19
meet_year                                  2022
meet_country                          Singapore
meet_name           Singapore Powerlifting Open
Name: 160540, dtype: object
In [95]:
# 172565
df3.iloc[172553]
Out[95]:
name             Muhammad Ghazi Bin Hishamuddin
sex                                           M
event                                       SBD
age_division                              Opens
bodyweight                                58.85
weight_class                                 59
best_squat                                152.5
best_bench                                115.0
best_deadlift                             182.5
total                                     450.0
place                                         3
dots_score                               385.96
birth_country                           Unknown
federation                                   PS
meet_date                            2022-12-10
meet_year                                  2022
meet_country                          Singapore
meet_name                 Ark Powerlifting Open
Name: 172553, dtype: object

It's important to acknowledge that maintaining a weight of 59kg at my height (167cm) may not have been optimal 😭 There is potential for further muscle development and optimization of my physique, which could have positively impacted my performance. But I'm curious to see where I stand among my U59 lifters!¶

In [96]:
plt.figure(figsize=(6, 6))
sns.violinplot(x='weight_class', y='total', data=df3[df3['weight_class'] == '59'])
plt.axhline(y=450, color='red', linestyle='--', label='My Highest Total')
plt.title('Total Lifted by U59 Lifters')
plt.xlabel('Weight Class')
plt.ylabel('Total Lifted')
plt.show()

I sit right above the third quartile, though it would seem the the most frequent value is around the 380-420kg total range. Let's take a look whether this holds true for DOTS value!¶

dots_score 385.96¶
In [97]:
# Calculate mean DOTS for juniors
junior_u59 = df3[(df3['age_division'] == 'Junior') & (df3['weight_class'] == '59') & (df3['dots_score'] != 0)]['dots_score'].mean()

# Calculate mean DOTS for opens
opens_u59 = df3[(df3['age_division'] == 'Opens') & (df3['weight_class'] == '59') & (df3['dots_score'] != 0)]['dots_score'].mean()

print("Mean DOTS for Junior:", junior_u59)
print("Mean DOTS for Opens:", opens_u59)

Mean DOTS for Junior: 358.9766091954022
Mean DOTS for Opens: 378.35168356997974

That's heartwarming to see that I'm slightly above average! (quite apparent with the others things I have in my life too 🤣😪)¶

Let's look through a scatter plot illustrating the relationship between total weight lifted and bodyweight, with points colored according to gender and a regression line overlay to discern the overall trend.¶

In [98]:
gender_palette = {'M': 'mediumblue', 'F': 'hotpink'}

plt.figure(figsize=(10, 6))
sns.scatterplot(data=df3, x='bodyweight', y='total', hue='sex', palette=gender_palette)
sns.regplot(x='bodyweight', y='total', data=df3, scatter=False, line_kws={'color': 'red'})

plt.show()


# R-squared calclulation
from sklearn.linear_model import LinearRegression
from sklearn.metrics import r2_score

x = np.array(df3['bodyweight'])  
y = np.array(df3['total'])      

# Model total regression
total_model = LinearRegression().fit(x.reshape(-1, 1), y)
total_r = r2_score(y, total_model.predict(x.reshape(-1, 1)))

print("R-squared value:", total_r)

R-squared value: 0.2887657518367255

The scatter plot shows a clear separation between male and female lifters based on their total scores and bodyweights, with male lifters generally achieving higher totals and having higher bodyweights. The line of points along the x-axis at total = 0 would likely represents lifters who were disqualified or did not complete any successful lifts.¶

For both male and female lifters, there is a positive linear correlation between bodyweight and total score. As bodyweight increases, the total score tends to increase as well, which aligns with the expectation that heavier lifters generally have the potential to lift more total weight! However, the spread of points increases with higher bodyweights, indicating greater variation in total scores among higher weight classes compared to lower weight classes.¶

While the overall trend is linear and positive, there is still substantial variation within each bodyweight range. The R-squared value of approximately 0.29 indicates that around 29% of the variability in total scores can be explained by bodyweight alone. This implies that a considerable portion of the variation in total scores is attributed to factors other than bodyweight, highlighting the complexity of performance in powerlifting competitions.¶

Let's take a look at where I stand in this scatter plot!¶

In [99]:
my_records = df3[df3['name'] == "Muhammad Ghazi Bin Hishamuddin"]

plt.figure(figsize=(10, 6))
sns.scatterplot(data=df3, x='bodyweight', y='total', hue='sex',  palette=gender_palette)

for idx, row in my_records.iterrows():
    plt.scatter(row['bodyweight'], row['total'], color='yellow', marker='*', s=100, label='Your Record' if idx == 0 else None)

plt.legend()
plt.show()

As discussed earlier; in comparison with the other male lifters around the same bodyweight, it would seem that I'm on the higher end of the distribution. Those 3 years of hard work did not put to waste!! However, there is still room for improvement as there are better performing lifters in my weightclass. 💪¶

Let's take a look at top level lifters in Singapore. One of them our very own Singapore Powerlifting (SGPL) president, Clinton Lee. Clinton competes in the 74 kg weight class and he has represented Singapore numerous times in his liftetime as a powerlifter.¶

image.png

Credits: https://www.instagram.com/clintoncratus/¶
In [100]:
clinton_lee = df3[df3['name'] == 'Clinton Lee #1']
clinton_lee_totals = df3[df3['name'] == 'Clinton Lee #1']['total']

#Violin Plot
plt.figure(figsize=(6, 6))
sns.violinplot(x='weight_class', y='total', data=df3[df3['weight_class'] == '74'])
plt.axhline(y=clinton_lee_totals.max(), color='red', linestyle='--', label='Clinton Highest Total')
plt.title('Total Lifted by U74 Lifters')
plt.xlabel('Weight Class')
plt.ylabel('Total Lifted')
plt.show()

# Scatter plot
plt.figure(figsize=(10, 6))
sns.scatterplot(data=df3, x='bodyweight', y='total', hue='sex',  palette=gender_palette)

for idx, row in clinton_lee.iterrows():
    plt.scatter(row['bodyweight'], row['total'], color='yellow', marker='*', s=100, label='Clinton Lee Record' if idx == 0 else None)

plt.legend()
plt.show()

Clinton falls near the upper end of the dense area of the distribution, representing a relatively high percentile of the total weight lifted within the U74 weight class. He is an elite level lifter with years of experience under his belt, though that doesn't make him less human. From the scatter plot, it would appear that he failed to qualify a total in one of the meets. Let's take a look at what meet that might be:¶

In [101]:
df3.loc[(df3['name'] == 'Clinton Lee #1') & (df3['total'] == 0)]
Out[101]:
name sex event age_division bodyweight weight_class best_squat best_bench best_deadlift total place dots_score birth_country federation meet_date meet_year meet_country meet_name
212510 Clinton Lee #1 M SBD Opens 73.4 74 262.5 170.0 0.0 0.0 DQ/NS 0.0 Singapore AsianPF 2023-12-11 2023 Malaysia Asian Classic Powerlifting Championships

It would appear that he did indeed missed all 3 attempts of his deadlift at the Asian Classic Powerlifting Championships hosted in Malaysia.¶

Thank you for representing Singapore for us Clinton! We wish him the best of luck for the upcoming World Open Classic Powerlifting Championships 2024, set to take place from 15-23 June 2024 in Druskininkai, Lithuania!¶

I'm intrigued to know the correlation between all the numeric columns. This could be an indiciation of what contributes most in a lifter's total score¶

In [102]:
numeric_columns = ['bodyweight', 'best_squat', 'best_bench', 'best_deadlift', 'total', 'dots_score']
correlation_matrix = df3[numeric_columns].corr()

plt.figure(figsize=(10, 4))
sns.heatmap(correlation_matrix, annot=True, cmap='coolwarm', fmt=".2f", annot_kws={"size": 10})
plt.title('Correlation Matrix')
plt.show()

It would seem that deadlift has a very strong correlation (0.92) with total, while bench has a smaller correlation (0.89).Both the squat and deadlift appear to be strong contributors to the total score, with the bench being slightly less influential but still positively correlated.¶

In [103]:
plt.figure(figsize=(10, 6))

# Plotting scatter plots with regression lines for each lift
sns.regplot(x='bodyweight', y='best_squat', data=df3, scatter_kws={'alpha': 0.1, 'color': 'lightgrey'}, line_kws={'color': 'red'})
sns.regplot(x='bodyweight', y='best_bench', data=df3, scatter_kws={'alpha': 0.1, 'color': 'lightgrey'}, line_kws={'color': 'blue'})
sns.regplot(x='bodyweight', y='best_deadlift', data=df3, scatter_kws={'alpha': 0.1, 'color': 'lightgrey'}, line_kws={'color': 'green'}
           )
# Dummy scatter plots for legend
plt.scatter([], [], color='green', label='Best Deadlift')
plt.scatter([], [], color='red', label='Best Squat')
plt.scatter([], [], color='blue', label='Best Bench')
plt.legend(loc='upper left')

plt.title('Regression Line for Best Squat/Bench/Deadlift')
plt.xlabel('Bodyweight (kg)')
plt.ylabel('Weight Lifted (kg)')
plt.legend()
plt.grid(True)
plt.show()


### R-squared calculation
x = np.array(df3['bodyweight'])  
y1 = np.array(df3['best_squat'])      
y2 = np.array(df3['best_bench'])      
y3 = np.array(df3['best_deadlift']) 

from sklearn.linear_model import LinearRegression
from sklearn.metrics import r2_score

# Squat regression
squat_model = LinearRegression().fit(x.reshape(-1, 1), y1)
squat_r = r2_score(y1, squat_model.predict(x.reshape(-1, 1)))

# Bench regression
bench_model = LinearRegression().fit(x.reshape(-1, 1), y2)
bench_r = r2_score(y2, bench_model.predict(x.reshape(-1, 1)))

# Deadlift regression
deadlift_model = LinearRegression().fit(x.reshape(-1, 1), y3)
deadlift_r = r2_score(y3, deadlift_model.predict(x.reshape(-1, 1)))

print("Squat R-squared value:", squat_r)
print("Bench R-squared value:", bench_r)
print("Deadlift R-squared value:", deadlift_r)

Squat R-squared value: 0.35087975513910763
Bench R-squared value: 0.3451292200969316
Deadlift R-squared value: 0.28280928588889953

This is very interesting observation! It is obvious that bench performance will always be weaker than the other two lifts. However, given that the squat regression line is steeper, this indicates that squat performance increases at a faster rate with increasing bodyweight compared to deadlift performance, based on the regression models shown! It would seem at heavier bodyweights, lifters tend to squat more than their deadlifts. We can take a look at Ray William's best performance:¶

In [104]:
df3.iloc[df3[df3['sex'] == 'M']['total'].idxmax()]
Out[104]:
name                     Ray Williams #1
sex                                    M
event                                SBD
age_division                       Opens
bodyweight                         190.4
weight_class                        120+
best_squat                         490.0
best_bench                         247.5
best_deadlift                      375.0
total                             1112.5
place                                  1
dots_score                        562.77
birth_country                        USA
federation                         USAPL
meet_date                     2019-03-02
meet_year                           2019
meet_country                         USA
meet_name        Arnold SBD Pro American
Name: 50656, dtype: object

Weighing at 190kg, he did infact squat (490.0) more than he deadlifted (375.0)! What a remarkable feat of strength.¶

To conclude, it's important to note that the R-squared values for these regression lines were modest, ranging from 0.28 for deadlift to 0.35 for squat. This suggests that while bodyweight is a contributing factor, there are likely other variables that significantly influence lift performance, which are not captured in these simple linear regression models.¶

In summary, this EDA highlights the importance of all three lifts in contributing to the total score, with deadlift potentially having the most substantial impact. Additionally, it suggests that higher bodyweight may lead to better squat performance relative to bench press and deadlift. However, further analysis is required to understand the complex interplay of variables that determine lift performance in powerlifting.¶